Climate risk starts with a physical process, but decisions depend on how that process reaches assets, cash flows, people, and ecosystems. This guide takes you from greenhouse forcing and carbon accounting through physical and transition risk, sustainable finance, scenario analysis, net zero, nature risk, and transition reporting. You will learn to trace a hazard or policy change into credit, market, liquidity, and operational effects; compare carbon metrics without confusing emissions with financial exposure; and test whether a target or financing claim has credible boundaries and evidence. Worked examples show how to calculate emissions and portfolio measures, interpret scenario assumptions, and choose actions when data are incomplete. The focus is applying these mechanisms to practical decisions.
Foundations of Climate Change: What Is Climate Change?
1. Earth's energy balance
Earth absorbs part of the sunlight reaching it and reflects the rest. To maintain a stable temperature over time, the climate system must emit roughly as much energy to space as it absorbs, mainly as infrared radiation. A persistent positive difference adds energy to the oceans, land, atmosphere, and ice. Oceans store most of the added heat, so surface air temperature need not rise smoothly each year. When interpreting an energy imbalance, check the direction of each flow and the period averaged. A short cool spell describes local or temporary conditions; it cannot settle whether the planet is accumulating energy. The 2026 SCR curriculum lists climate foundations among its topics, and this balance is the starting point for tracing later physical risks.
Worked example: Suppose global average absorbed sunlight is 240 W/m² and outgoing radiation is 239 W/m² over a sustained period. The net is 240 minus 239, or positive 1 W/m². That means the climate system gains one joule per second for each square meter represented by the global average. A cool local season may coexist with that persistent global gain.
Watch the distinction: A common error is to treat one year's air temperature as a complete energy balance. Heat can move among the atmosphere, oceans, land, and ice while the system keeps gaining energy.
Reference: [1]
2. Greenhouse forcing and feedback
A forcing changes the climate system's energy balance before the resulting temperature response. Adding CO2 increases the atmosphere's absorption of outgoing infrared radiation, creating an initial warming influence. A feedback follows that temperature change and alters its size. Warmer air can contain more water vapor; because water vapor also absorbs infrared radiation, this generally amplifies warming. Loss of reflective snow or ice exposes darker surfaces that absorb more sunlight, another amplifying feedback. Some responses, including aspects of cloud change, are harder to pin down and can differ by region. For a causal question, identify the first external change, then ask which processes respond to warming. This distinction matters because reducing an initial forcing and estimating the eventual temperature response are related but separate tasks.
Worked example: Suppose extra CO2 reduces outgoing infrared energy at first. Temperature then rises, allowing more atmospheric water vapor and melting some reflective ice. The CO2 increase is the initiating forcing. The added water vapor increases infrared absorption, while the darker surface increases solar absorption. Both reinforce the response; neither is the original CO2 addition.
Watch the distinction: Calling water vapor the initial cause because it is a greenhouse gas reverses this sequence. In this example, its concentration rises in response to warming initiated by added CO2.
Reference: [1]
3. Carbon stocks, flows, and budgets
Atmospheric CO2 is a stock, while emissions are a flow into the wider carbon cycle. Natural sinks remove some emitted CO2, but a substantial fraction persists long enough that repeated annual emissions raise the atmospheric stock. A remaining carbon budget expresses an estimated cumulative amount of future net CO2 emissions associated with a specified temperature outcome and likelihood. It is not a calendar date built into the atmosphere. To use one, add projected annual net emissions, including any credible removals, until the cumulative total reaches the stated budget. Check the temperature objective, probability, starting date, treatment of other gases, and scientific uncertainty before comparing estimates. Cutting emissions sooner lowers cumulative additions; keeping emissions flat uses the budget at a constant rate even if the annual figure stops growing.
Worked example: Assume a hypothetical remaining budget of 100 GtCO2 and constant net emissions of 10 GtCO2 per year. Dividing 100 by 10 gives ten years to exhaust that assumed budget. If emissions instead decline, cumulative use grows more slowly and the budget lasts longer. The calculation is conditional on the inputs and cannot guarantee a ten-year deadline.
Watch the distinction: Do not confuse stabilizing annual emissions with stabilizing atmospheric CO2. A continuing positive net emissions flow keeps adding to the cumulative total, even when the yearly flow stays constant.
Reference: [1]
4. Comparing greenhouse gases
Greenhouse gases differ in how strongly they absorb radiation and how long their effects persist. Carbon dioxide equivalent, written CO2e, converts a quantity of another gas into the amount of CO2 assigned an equivalent warming effect under a specified metric and time horizon. The usual calculation multiplies tonnes of gas by its chosen conversion factor. This allows emissions from different gases to be added within one inventory, but the resulting total inherits the metric's assumptions. A short-lived gas can look different under a shorter horizon than under a longer one. Always identify whether a question gives mass of methane itself or mass already expressed as CO2e, and do not multiply a CO2e figure again. Comparisons also require consistent reporting boundaries and units; a precise conversion cannot repair an incomplete inventory.
Worked example: For a hypothetical exercise, assign methane a 100-year conversion factor of 28 tonnes CO2e per tonne of methane. Two tonnes of methane then equal 2 × 28 = 56 tonnes CO2e on that assumed basis. Adding 10 tonnes of CO2 would give 66 tonnes CO2e. A different time horizon or factor would change the methane result.
Watch the distinction: A CO2e number is not the physical mass of CO2 released. It is a comparison measure tied to a chosen factor and time horizon, which must travel with the result.
Reference: [1]
5. Weather, climate, and attribution
Weather describes conditions over short periods and particular places; climate describes their statistical pattern over longer periods. An isolated cold day can occur in a warming climate because seasonal variation and circulation still operate. Event attribution asks a narrower question: how much did human-driven climate change alter the likelihood or severity of a specified type of event? Analysts compare a modeled world with the relevant human influence against a counterfactual world without it, while accounting for uncertainty and the event definition. A risk ratio compares the two event probabilities; an intensity comparison asks how much hotter, wetter, or stronger the event became. Neither method generally assigns a single physical event to one exclusive cause. Check the location, threshold, time period, and uncertainty before translating an attribution result into an asset decision.
Worked example: Suppose a defined heatwave has a modeled annual probability of 1% without the added human influence and 2% with it. The risk ratio is 2% ÷ 1% = 2, so the event is estimated to be twice as likely under those assumptions. It may still arise from several interacting weather conditions; the estimate does not identify one sole cause.
Watch the distinction: One cold winter day cannot disprove a multidecade warming trend. It samples short-term weather, while the trend and attribution estimate concern distributions across many places or years.
Reference: [1]
6. Warming and physical consequences
Warming changes several connected physical processes. Hotter air raises direct heat exposure and can increase evaporation; warmer air can also hold more moisture, contributing to heavier rainfall when conditions release it. Ocean water expands as it warms, while melting land ice adds water to the sea. These processes raise mean sea level, changing the starting point from which tides and storm surges act. Local outcomes still depend on terrain, drainage, defenses, coastal subsidence, and how people use the site. For an asset assessment, separate the broad climate driver from the local hazard and then from the asset's vulnerability. A seawall designed from historical surge height alone may miss changes in baseline sea level, wave action, or the chance of overtopping over its service life.
Worked example: Consider a hypothetical port whose historical extreme surge reached 2.0 meters above an old mean sea level. If local mean sea level later rises 0.4 meter and a similar surge occurs, the water surface would reach about 2.4 meters relative to that old reference. A wall built only to 2.0 meters lacks 0.4 meter before considering waves or safety margin.
Watch the distinction: Do not treat a global temperature increase as a direct forecast of one site's flood depth. Local elevation, land movement, storms, and defenses determine how a broad change reaches an asset.
Reference: [1]
Sustainability
7. Two views of materiality
An organization can examine sustainability through two directions of effect. Financial materiality asks how environmental or social conditions could change its cash flows, financing, or value. Impact materiality asks how its activities affect people and ecosystems, whether or not the effect has already reached its accounts. These directions can interact: heavy water use may harm a community first and later expose the user to shortages, restrictions, or conflict. To analyze a case, state the affected party, describe the causal path, and identify the time horizon. Avoid assuming that every important impact has an immediate monetary estimate or that every financial exposure began with the firm's own conduct. Reporting frameworks can set different formal requirements, so the underlying distinction is more durable than any one disclosure rule.
Worked example: A factory withdraws enough river water to strain a nearby town's supply. That effect on residents and the river matters even while factory profits are steady. Later, lower river flow forces two weeks of reduced production, cutting earnings. The water issue now has a financial channel too. The two assessments concern different directions of effect, not competing descriptions.
Watch the distinction: An issue does not become relevant to people only when it affects profit. Conversely, a financial climate exposure can arise from conditions the organization did not itself create.
Reference: [1]
8. Sustainability trade-offs
A project can improve one sustainability outcome while worsening another. Avoided fossil-fuel generation is useful evidence for a solar project, but it does not settle land, water, habitat, or community effects. Begin by defining the decision and credible alternatives, including a different location or design. Compare the expected benefits and harms over the project's life, identify who bears each effect, and check whether mitigation is feasible and funded. An impact that falls on a particular community cannot be dismissed merely because an aggregate emissions figure looks favorable. Likewise, a concern should be tested against evidence rather than assumed to cancel every climate benefit. The practical question is which feasible option delivers the intended energy service with acceptable environmental and social consequences, and what monitoring would reveal if assumptions fail.
Worked example: A proposed solar site is expected to displace fossil generation but would occupy productive farmland and use scarce local water for cleaning. A second site has slightly lower output but avoids both pressures. Decision makers should compare lifetime generation, emissions displaced, land effects, water demand, cost, and affected households before choosing. The first site's climate benefit alone cannot establish that it is preferable.
Watch the distinction: A green label describes an intended benefit, not a complete impact assessment. It can conceal trade-offs when the project boundary excludes land use, water, or affected people.
Reference: [1]
9. Responsible banking: alignment and impact
The first Principle for Responsible Banking concerns aligning a bank's strategy with societal goals; the second concerns identifying significant impacts, setting targets, and pursuing them. Alignment requires more than naming a broad goal. A bank must understand where its products and financing shape outcomes, then choose priorities that fit its portfolio and operating context. Impact analysis considers positive and negative effects and identifies the areas where the bank can make a meaningful difference. Targets need a defined baseline, measure, scope, and timeframe so that progress can be tested. Actions might include lending criteria, client support, and capital allocation. The SCR required readings include the responsible banking principles; treat them as a framework for bank practice, not as a claim that a pledge itself achieves an outcome.
Worked example: Suppose a bank finds that commercial-property loans are its largest source of financed building emissions. It records a portfolio baseline, sets a measurable reduction target for that segment, and offers financing tied to credible building upgrades. It then tracks completed upgrades and portfolio emissions. A general promise to support sustainability would neither identify the main impact nor show how lending changes it.
Watch the distinction: Choosing a convenient small activity can produce an attractive target while missing the bank's most significant impact. Start with the impact assessment, then set the target and actions.
10. Responsible banking: clients and stakeholders
The third banking principle focuses on working with clients and customers to encourage sustainable practices; the fourth brings relevant stakeholders into decisions. These are different sources of information. Borrowers can explain their assets, constraints, financing needs, and feasible transition steps. Workers, tenants, communities, and other affected groups can identify consequences that a borrower or bank may overlook. A bank can use client engagement to make lending conditions and support practical, while stakeholder input tests whether the proposed change shifts costs or harms elsewhere. The sequence is to identify the bank's material impact, engage the parties who can affect or experience it, alter the plan where evidence warrants, and monitor results. Consultation alone is not proof of a good outcome; its value lies in how it informs decisions.
Worked example: A bank finances insulation and heating upgrades for rental buildings. It asks borrowers for installation schedules, projected energy savings, and maintenance plans. Tenant consultation reveals that proposed rent increases could offset lower energy bills. The bank can compare net household costs and adjust financing or project terms. Borrower engagement improves execution; tenant input exposes a social impact requiring a response.
Watch the distinction: Speaking only with the borrower does not necessarily cover everyone affected by a loan. The borrower's transition plan and tenants' affordability concerns answer different questions.
11. Responsible banking: governance and accountability
The fifth banking principle connects implementation to governance and culture; the sixth concerns transparency and accountability. Governance turns a target into assigned decisions: the board oversees direction and risk appetite, senior managers allocate resources, and business teams apply changes in lending and customer work. Incentives and staff capability affect whether those responsibilities operate in practice. Transparency requires reporting the chosen measures, progress, limitations, and missed milestones clearly enough for others to evaluate performance. Accountability means explaining deviations and corrective action rather than quietly changing a baseline or excluding difficult exposures. A sound review follows the chain from target to owner, budget, operational control, measured result, and public explanation. A report can be polished while implementation remains weak, so evidence of actual decisions matters alongside disclosure.
Worked example: A bank gives its board oversight of a property-lending emissions target, assigns managers a retrofit-financing budget, and includes progress in regular risk reviews. At year-end it misses a milestone because upgrades ran late. It reports the shortfall, the affected loan segment, and revised actions. Ownership and candid progress reporting make the target more accountable than publication alone.
Watch the distinction: Publishing a target is not the same as governing it. Without decision owners, resources, performance checks, and explanation of missed results, the commitment has little operational force.
12. The six responsible investment principles
The six Principles for Responsible Investment translate environmental, social, and governance considerations into investor practice. They call for incorporating ESG issues into investment analysis and decisions; acting as an owner who includes ESG in ownership policies; seeking appropriate ESG disclosure from investees; promoting adoption of the principles across the investment industry; working with others to improve effectiveness; and reporting activities and progress. These are commitments to processes and conduct, not a guarantee that every holding has low emissions or high performance. In a case, map the investor's action to the correct principle: security analysis differs from voting, requests to issuers differ from the investor's own reporting, and collaboration differs from promotion to other market participants. Implementation quality depends on whether these actions influence real investment and ownership decisions.
Worked example: An investor adds transition risk to valuation models, votes on an issuer's transition plan, requests emissions data, asks external managers to adopt responsible investment practices, joins peers in engagement, and publishes its own progress. Those actions correspond in order to the six principles. Its reported activity should still show what changed; listing the principles alone would not demonstrate implementation.
Watch the distinction: Do not collapse all six principles into buying a sustainability fund. They also address ownership, investee disclosure, industry adoption, collaboration, and the investor's own reporting.
Climate Change Risk
13. Acute and chronic physical risk
Physical climate risk can arise from acute events, such as floods or storms, or chronic shifts, such as rising average heat or changing rainfall patterns. The distinction guides the evidence required. Acute analysis examines event likelihood and severity, site elevation, defenses, and recovery time. Chronic analysis examines trends, operating thresholds, resource availability, maintenance, and adaptation over the asset's life. A facility can face both: hotter conditions may raise ordinary cooling costs while a severe storm causes a sudden closure. Start with the relevant hazard, locate the asset, then identify which features make it vulnerable and which controls reduce loss. Historic experience can help describe present exposure but may understate future conditions if the hazard distribution changes. A broad hazard map alone cannot show the asset's exact loss.
Worked example: A warehouse beside a river faces a one-day flood and progressively hotter summers. For flood risk, assess possible depth at its floor, inventory placement, and time needed to reopen. For heat, estimate rising cooling demand and worker exposure across many summers. A flood barrier addresses the acute event; ventilation and work practices address the chronic pressure.
Watch the distinction: A sudden event and a slow trend require different data and controls. Labeling both simply 'physical risk' can hide the timing and mechanism needed for a decision.
Reference: [1]
14. Transition-risk drivers
Transition risk arises as the economy changes in response to climate goals, preferences, or innovation. Policy can change taxes, standards, or permitted activity. Technology can make an existing process less competitive. Markets can shift demand, supply, input costs, or access to capital. These drivers may interact: a new rule can accelerate adoption of a cheaper substitute, while changing customer demand can erode revenue before regulation arrives. Identify the first driver, trace it through volume, prices, costs, or asset life, and then assess whether the firm can adapt. A stranded asset is one whose expected economic value is impaired before its planned end of use. The classification does not require physical damage. Timing matters because an asset with large fixed costs may lose value when expected future cash flows change, even if current operations continue.
Worked example: A gas plant expects strong evening electricity revenue for another decade. Lower-cost storage enters the market and supplies that demand, reducing the plant's projected operating hours and cash flow. No carbon price changes and the plant is undamaged. The immediate driver is technology, transmitted through market competition; the resulting loss can impair the plant's value before its planned retirement.
Watch the distinction: Transition risk is broader than carbon taxes. A technology or demand shift can weaken an asset's economics even when no new climate policy directly applies to it.
Reference: [1]
15. Legal and reputation channels
Climate exposure can reach a firm through legal claims and reputation as well as through physical hazards or transition costs. A contested disclosure may lead to investigation, litigation expense, remediation, or changed sales. Reputation effects arise when customers, employees, lenders, or investors respond to perceived conduct, even before a legal outcome. To analyze the channels, identify the claim or action, the evidence supporting it, the parties likely to respond, and the financial transmission path. Do not automatically count every potential response as a separate loss: lower sales and a valuation decline may reflect the same expected earnings reduction. Also distinguish a challenged claim from a proven violation. The risk estimate should describe plausible consequences and uncertainty, while the underlying evidence should separate gross emissions, credits, and the boundaries of any advertised environmental benefit.
Worked example: A firm calls a product carbon neutral but cannot document the credits used to support that claim. A challenge causes legal fees and prompts some customers to switch brands. The legal expense is one cash outflow; lost sales reduce operating income. Neither requires flood or heat damage. The assessment should avoid adding a share-price decline if it merely capitalizes those same expected losses.
Watch the distinction: A climate-related loss need not originate in damaged premises. Unsupported claims can trigger legal costs and stakeholder responses, but allegations should not be presented as established violations.
Reference: [1]
16. Climate effects on credit risk
Credit risk is the chance that a borrower fails to meet its obligations and the loss a lender suffers if it does. A simple expected credit loss estimate multiplies exposure at default, probability of default, and loss given default for a defined period. Climate shocks can affect probability of default through weaker cash flow or higher costs; they can affect loss given default if collateral becomes harder to sell or less valuable. The same shock may also change exposure if a borrower draws a credit line. Keep assumptions consistent with the period and scenario, and avoid treating a single expected loss as the full range of outcomes. For related risk techniques, the financial risk manager guide offers a useful comparison of credit, market, liquidity, and operational channels.
Worked example: Harbor Bank has a hypothetical $20 million loan exposure with 40% loss given default. Before a climate shock, a 2% default probability gives $20 million × 0.02 × 0.40 = $160,000 expected loss. If the probability rises to 5% while other inputs stay fixed, expected loss becomes $400,000. The increase is $240,000, driven here by weaker borrower credit quality.
Watch the distinction: Do not equate expected loss with the amount that will certainly be lost. It is an average under stated probability and recovery assumptions, and severe outcomes can be much larger.
17. Market and liquidity risk
Market risk is loss from changes in assessed asset prices or rates. Liquidity risk appears when a position cannot be sold promptly near that assessed value, or funding cannot be obtained on acceptable terms. A climate policy surprise may lower a bond's expected cash flows and therefore its fair value. If many holders then try to sell, thin trading can widen the gap between quoted value and executable price. When breaking down a sale, compare the original marked value with the new assessed value for the market move, then compare the new assessed value with actual execution for the sale-related liquidity cost. The split depends on sound valuation and timing; it is an analytical decomposition, not a claim that markets reveal two perfectly separate causes.
Worked example: Harbor Bank holds one hypothetical bond marked at 100 before a policy announcement. Its assessed price falls to 92, an eight-point market loss. The bank must sell quickly and receives 91, one point below the new assessed price. Total realized decline is nine points: eight from repricing and one attributed to execution in a thin market.
Watch the distinction: Calling the entire nine-point decline liquidity loss hides the policy-driven repricing. Calling it entirely market loss hides the extra cost of an urgent sale at 91.
18. Operational disruption and resilience
Operational risk includes losses from failed processes, systems, people, or external events. A flood can interrupt a payment center even if its loan portfolio and investments remain sound. Estimate the relevant downtime, the cost per period, additional recovery expenses, and any obligations that continue during closure. Then test controls against the specific failure: a backup site helps only if it can take over the required processing volume, staff can reach it, and data and communications work. A resilience measure can lower loss severity or duration without changing the flood hazard itself. Compare the modeled outcome with and without the control, and retain residual risks such as simultaneous disruption at both sites. A business continuity plan is stronger when exercised under realistic conditions than when it exists only as documentation.
Worked example: Harbor Bank estimates that flooding its payment center would stop service for five days at a hypothetical $200,000 loss per day. The modeled interruption loss is 5 × $200,000 = $1 million. A tested backup reduces closure to one day, or $200,000. The modeled reduction is $800,000, assuming the daily cost and backup performance hold.
Watch the distinction: A backup does not eliminate the flood or guarantee zero loss. It changes the expected interruption path only if capacity, staff access, data, and communications remain usable.
Sustainability and Climate Policy, Culture, and Governance
19. Climate goals and national policy
International climate goals establish a shared direction, while national commitments express what individual countries intend to contribute. Domestic laws, budgets, infrastructure programs, and sector measures determine whether that intention changes emissions and resilience in practice. A pledge is therefore an input to analysis, not evidence of delivery. For a financial decision, trace the sequence from stated goal to adopted measure, implementation timetable, affected activity, and observed outcome. A planned rule may shape expectations before it takes effect, yet its probability, design, and enforcement can remain uncertain. Comparing countries also requires care because targets may use different baseline years, sectors, gases, and accounting methods. The most useful distinction is between ambition, policy implementation, and measured results, especially when valuing assets with long lives.
Worked example: A country announces a lower-emissions target, funds grid upgrades, and adopts rules for electricity generators. The target describes ambition; the appropriated grid spending and enacted rules are implementation steps. If emissions then decline in the covered sector, observed data provide evidence of progress. A lender should still examine whether its borrower's plant faces the rules during the loan term.
Watch the distinction: Do not price a national pledge as though it were already fully implemented. Its financial effect depends on later measures, timing, enforcement, and the activities covered.
Reference: [1]
20. Carbon-price decisions
A carbon price gives an emitting activity a monetary cost per tonne, while abatement reduces the tonnes exposed to that cost. For a simple decision, compare the marginal cost of avoiding each tonne with the price paid for emitting it. If abatement costs less per tonne and works as assumed, it lowers total cost; if it costs more, paying the price may be cheaper in the short run. Real decisions also depend on capital spending, timing, technology limits, changing prices, and whether costs can be passed to customers. Specify whether the price applies to all remaining emissions and whether abatement costs are annual or one-time before calculating. A correct arithmetic answer is only as good as those assumptions. SCR practice questions can help reinforce this cost comparison.
Worked example: Assume a hypothetical carbon price of $50 per tonne and annual emissions of 1,000 tonnes. Paying for all emissions costs 1,000 × $50 = $50,000. Abating 600 tonnes at $30 each costs $18,000; pricing the remaining 400 costs $20,000. Total annual cost is $38,000, a $12,000 saving, assuming the abatement cost and price apply for the same year.
Watch the distinction: Comparing the $18,000 abatement bill with the original $50,000 carbon bill alone omits the price on remaining emissions. Add both costs before judging the option.
Reference: [1]
21. Six NGFS recommendations
The six NGFS recommendations in its 2019 framework form a sequence for making climate risk visible and manageable in finance. Central banks and supervisors should integrate climate risk into financial stability monitoring and supervision, while considering sustainability in their own portfolio management. They should help close data gaps, build technical capacity and share knowledge. The remaining recommendations support consistent climate and environmental disclosure and the development of taxonomies that classify economic activities. Better data and disclosure let institutions compare exposures; capacity lets staff interpret those comparisons; supervision turns the findings into risk decisions. A taxonomy can make classifications clearer but cannot measure a bank's loss on its own. These recommendations guide practice; their presence in the SCR required readings does not make them current legal requirements everywhere.
Worked example: A supervisor finds that banks report coastal property exposures using incompatible location data. It asks for more consistent disclosure, trains examiners to assess flood assumptions, and examines concentrations during supervision. Its central bank separately reviews climate exposure in its own holdings. A shared taxonomy helps classify financed activities, but site-level hazard and borrower information are still needed to estimate losses.
Watch the distinction: Memorizing six labels without their users is misleading. Reviewing a central bank's own investments differs from supervising banks, and a taxonomy cannot replace exposure measurement.
22. Board oversight and risk ownership
Climate risk reaches governance through ordinary business decisions: where to lend, what collateral to accept, how to price risk, and when to reduce a concentration. The board sets strategy and risk appetite, receives material risk information, and holds management accountable. Business units own the risks they take and must apply controls in origination and monitoring. An independent risk function challenges assumptions, aggregates exposures, tests limits, and escalates breaches. Internal audit provides independent assurance that governance and controls operate as designed. Keep decision rights clear: risk staff can challenge a proposed loan without becoming its commercial sponsor, while audit should test the process rather than approve individual deals. Useful board information shows exposure, uncertainty, limit use, and actions, not just a climate score.
Worked example: A lending team proposes a coastal hotel loan and estimates that insurance will cover flood damage. Risk staff test the policy exclusions, flood depth assumptions, and refinancing exposure. The board has set a limit for vulnerable collateral, so management must show whether this loan fits it. Later, audit checks that the approved flood review and exception process were actually followed.
Watch the distinction: Sending a climate report to the board does not transfer loan ownership to directors. The lending unit remains responsible for its decision and controls.
Reference: [1]
23. GHG accounting principles
A useful GHG inventory follows five accounting principles. Relevance means its boundary and detail serve the decisions of intended users. Completeness calls for all material sources within that boundary, with exclusions identified. Consistency preserves comparability across periods or explains changes in methods and boundaries. Transparency makes assumptions, factors, exclusions, and revisions traceable. Accuracy calls for reasonable estimates and efforts to reduce material bias; it does not require impossible precision. Apply the principles together. A precise meter reading from one factory cannot rescue an inventory that omits another material site. Likewise, an estimate can be acceptable when its method, uncertainty, and improvement plan are disclosed. When a boundary changes, compare like with like before interpreting an emissions trend as operational progress or deterioration.
Worked example: A company reported 8,000 tCO2e last year but omitted a leased facility that produced an estimated 1,000 tCO2e. This year it includes the site and reports 8,500 tCO2e. If other methods are comparable, a restated prior figure of 9,000 tCO2e shows a 500 tCO2e decrease, about 5.6%, rather than an apparent 500 tCO2e increase.
Watch the distinction: A higher reported total need not mean worse operations. First check whether sites, calculation methods, and emission factors changed between the two years.
24. Organizational and operational boundaries
First choose an organizational boundary: the equity-share approach attributes emissions according to ownership interest, while a control approach consolidates operations under financial or operational control, depending on the chosen control criterion. Apply that choice consistently to ventures and subsidiaries. Then set operational boundaries for the included operations. Scope 1 covers emissions from sources they own or control, such as fuel burned on site. Scope 2 covers emissions from generating purchased electricity, heat, steam, or cooling they consume. Scope 3 addresses other value-chain emissions and is distinct from the direct-versus-purchased-energy classification. Boundary choice can change the reported total without changing physical emissions, so comparisons need the same approach. State the approach, exclusions, and treatment of joint ventures before interpreting a portfolio or company total.
Worked example: A firm owns 60% of an uncontrolled venture that emits 1,000 tCO2e directly. Under equity share it attributes 600 tCO2e. Under its chosen operational-control approach it excludes the venture's direct emissions from consolidated Scope 1. At a separate controlled site, fuel combustion is Scope 1; purchased electricity is Scope 2. The two approaches answer different consolidation questions.
Watch the distinction: Ownership of most shares does not automatically establish operational control. Nor does purchasing electricity make the generating plant's emissions direct emissions of the buyer.
Green and Sustainable Finance: Markets and Instruments
25. Use-of-proceeds green bonds
A use-of-proceeds green bond raises debt for stated eligible environmental projects. Assess two separate claims. The financing claim concerns whether proceeds are allocated to eligible spending, tracked while unallocated, and reported with enough detail to test the allocation. The credit claim concerns whether the issuer can pay coupons and principal under the bond terms. Green project benefits do not automatically improve the issuer's ability to repay, especially for a general obligation bond. Check the eligibility criteria, the allocation period, treatment of refinancing, and whether reported outcomes are estimates or measured results. Investors should also ask who verifies the allocation and what happens if a project changes. The ESG investing guide offers related context for assessing labeled instruments.
Worked example: An issuer sells a hypothetical $100 million green bond and records $60 million for solar construction and $40 million for grid upgrades. The allocation totals $100 million, so the proceeds can be reconciled to the stated uses. An investor still tests the issuer's cash flows and debt obligations. A project being green does not guarantee bond repayment or prove the project's eventual emissions savings.
Watch the distinction: A green label does not make the bond secured by green assets. Read the repayment terms separately from the proceeds allocation and impact report.
Reference: [1]
26. Sustainability-linked financing
Sustainability-linked financing changes a financial term when an issuer or borrower meets or misses a defined performance target. Proceeds can fund general activities; the link lies in the contractual performance mechanism, not a restricted project list. Read the metric, baseline, target date, observation method, verification requirement, and exact change in coupon or margin. A target may be weak if it reflects business as usual, covers too little of the borrower's impact, or can be met through accounting changes alone. A small pricing adjustment might provide limited incentive relative to the cost of genuine change. Also distinguish performance from disclosure: a borrower can report a metric accurately and still miss the target. Investors should evaluate both the credibility of the target and the borrower's underlying credit risk.
Worked example: A hypothetical $50 million loan increases its annual interest margin by 25 basis points if a verified emissions target is missed. Since 25 basis points equals 0.0025, the extra annual interest is $50 million × 0.0025 = $125,000 while the higher margin applies. The borrower may use loan proceeds generally; the target affects price, not permitted spending.
Watch the distinction: Treating this as a green use-of-proceeds loan confuses two structures. The decisive contract term is the measured target and its pricing consequence.
Reference: [1]
27. Project-finance resilience
Project finance depends heavily on cash generated by a specific asset, so climate disruption can quickly affect debt service. Debt service coverage ratio equals cash available for debt service divided by scheduled principal and interest for the same period. Start with operating revenue, subtract operating costs, taxes, and necessary spending under the financing definition, then stress the resulting cash. A flood may stop production, damage equipment, raise insurance costs, and delay repairs; avoid counting the same effect twice. Coverage below one means modeled cash is insufficient for that period, although reserves or support may temporarily bridge the gap. Test the timing of the shock, repair period, insurance exclusions, and any covenant or refinancing consequences. A single average-year ratio can conceal a severe short-term liquidity need.
Worked example: A hypothetical project has $12 million cash available for annual debt service and owes $10 million, giving coverage of 12 ÷ 10 = 1.20. Flood interruption reduces cash by $3 million, leaving $9 million and coverage of 9 ÷ 10 = 0.90. The $1 million cash shortfall requires a reserve, support, or another remedy; the ratio alone does not identify which is available.
Watch the distinction: A project can look profitable over its life yet miss a payment after a short disruption. Match stressed cash with debt due in that period.
Reference: [1]
28. Blended finance and first loss
Blended finance combines capital with different risk and return positions, often to support projects that otherwise struggle to attract private investment. In a simple first-loss structure, junior capital absorbs losses before senior capital. Work through the waterfall in order: identify total loss, subtract it from junior capital up to that layer's limit, then allocate any excess to senior investors. The junior layer reduces senior downside for losses within its capacity, but it does not make the project risk-free. Its provider still bears a real economic loss. Assess the size and terms of each layer, whether support is funded or merely promised, and whether expected project benefits justify the concession. A larger senior investment can be mobilized without proving that the project would never have happened otherwise.
Worked example: Suppose a hypothetical vehicle has $20 million junior capital and $80 million senior capital. A $15 million loss consumes $15 million of junior capital, leaving $5 million junior and all $80 million senior. If instead the loss were $25 million, junior capital would lose $20 million and senior capital $5 million. The waterfall changes loss allocation, not the project's total loss.
Watch the distinction: First-loss support protects senior investors only until the junior layer is exhausted. Describing it as a guarantee overstates the protection.
Reference: [1]
29. Taxonomy eligibility and alignment
Taxonomy eligibility asks whether an economic activity is covered by a classification system. Alignment asks whether the activity meets that system's applicable technical criteria and safeguards. Begin with the specific taxonomy and activity definition; then test the relevant performance conditions, any requirements to avoid significant harm to other objectives, and the prescribed social safeguards. The exact tests depend on the taxonomy and its version. An activity can be eligible but fail alignment, or lack the evidence needed to demonstrate alignment. That distinction matters when comparing investment products or estimating aligned revenue and capital expenditure. Do not infer that an entire company is aligned from one qualifying project. Record the denominator and the share being measured, and keep unsupported activity separate from activity shown to pass.
Worked example: A solar installation falls within the activities covered by a hypothetical taxonomy, so it is eligible. The investor has output data but no evidence for the applicable land-use safeguard or technical test. It can report the activity as eligible, subject to the taxonomy's definitions, but cannot count it as demonstrated aligned until those conditions are checked.
Watch the distinction: Eligible means the taxonomy has a category for the activity. It does not mean the activity has passed every criterion or that its issuer is aligned.
Reference: [1]
30. Additionality and measured impact
Measured impact needs a baseline that represents what would plausibly have happened without the intervention. Observed before-and-after change is useful but can include changes caused by demand, regulation, weather, or technology unrelated to the project. To estimate additional reduction, subtract actual emissions from a defensible no-project counterfactual over the same period and boundary. Explain how that counterfactual was chosen and test sensitivity to reasonable alternatives. A project may still provide value when additionality is uncertain, but its claimed incremental effect should reflect that uncertainty. Watch for leakage, where activity moves elsewhere, and for double counting, where multiple parties claim the same reduction. For financing decisions, separate project outcome from the additional effect of a particular investor's capital.
Worked example: A plant's emissions fall from 100,000 to 80,000 tCO2e, an observed drop of 20,000 tCO2e. Credible evidence suggests emissions would have fallen to 90,000 tCO2e without the project. Estimated additional reduction is therefore 90,000 − 80,000 = 10,000 tCO2e. If the counterfactual were lower, the additional estimate would shrink; the observed drop would remain 20,000.
Watch the distinction: Calling the whole before-and-after decrease additional assumes nothing else changed. A credible no-project path is needed before attributing the reduction to the intervention.
Reference: [1]
Climate Risk Measurement and Management
31. Hazard, exposure, and vulnerability
Physical risk combines a hazard, something exposed to it, and vulnerability to damage. For a simple event, expected annual damage equals annual event probability multiplied by exposed value and the fraction lost if the event occurs. The hazard estimate needs a location and time period; exposure needs asset value or cash flow at risk; vulnerability depends on construction, defenses, and operations. A changing climate can alter event probability and severity, so historical frequency may be a poor future estimate. One expected-loss number also hides tail risk: a lender may care about a rare loss that exceeds collateral or interrupts repayment. Consider indirect effects such as supply disruption separately and check for overlap. The financial risk manager guide covers related loss and risk concepts.
Worked example: A hypothetical warehouse worth $5 million faces a 2% annual flood chance and would lose 40% of its value if flooded. Expected annual direct damage is 0.02 × $5 million × 0.40 = $40,000. The conditional flood loss is $2 million. The $40,000 average does not mean losses arrive evenly each year or include business interruption.
Watch the distinction: A high hazard score alone does not specify financial loss. Asset location, value, defenses, and damage severity determine how the hazard becomes exposure.
Reference: [1]
32. From transition cost to financial ratios
Translate transition risk through the borrower's income statement before judging credit capacity. A carbon cost, fuel shift, or compliance investment first changes operating costs; pass-through may recover some of that change through prices, subject to demand and contracts. Net unrecovered cost reduces EBITDA if volumes and other costs stay constant. Interest coverage then compares EBITDA with interest expense, while debt service coverage may also include scheduled principal and cash adjustments. State which ratio is used. A lower ratio signals less room for further shocks but is not itself a default probability. Test whether pass-through reduces sales volume, whether costs persist, and whether capital spending or refinancing adds pressure. Comparing several assumptions is more informative than applying the full gross cost directly to earnings.
Worked example: A hypothetical borrower has $12 million EBITDA and $4 million annual interest, so EBITDA interest coverage is 12 ÷ 4 = 3.0. Transition costs rise $4 million, but price increases recover $2 million with unchanged volume. EBITDA becomes $10 million and coverage is 10 ÷ 4 = 2.5. The $2 million net cost, not the $4 million gross cost, drives this simplified result.
Watch the distinction: Passing costs to customers is not always costless. If higher prices reduce volume, the earnings effect can exceed the unrecovered cost in a fixed-volume calculation.
Reference: [1]
33. Weighted average carbon intensity
Weighted average carbon intensity, or WACI, combines each holding's share of portfolio value with that issuer's emissions per unit of revenue. Calculate issuer intensity using a consistent emissions scope, currency, revenue period, and unit; then multiply each intensity by its portfolio weight and sum. WACI helps compare the revenue-related emissions exposure of portfolios of different sizes. It does not attribute a number of emitted tonnes to the investor. A change can arise because holdings move, issuer emissions move, or revenue changes, including through prices or exchange rates. Check coverage and estimate quality before comparing portfolios: omitting high-emitting issuers can make a reported average look deceptively low. Use the metric alongside absolute and financed-emissions measures to understand what changed.
Worked example: A hypothetical portfolio invests 60% in issuer A at 200 tCO2e per $1 million revenue and 40% in issuer B at 50 tCO2e per $1 million revenue. WACI is 0.60 × 200 + 0.40 × 50 = 120 + 20 = 140 tCO2e per $1 million revenue. It is an average intensity, not 140 tonnes of financed emissions.
Watch the distinction: Multiplying issuer intensity by ownership share rather than portfolio weight calculates a different quantity. WACI weights by the investment's share of portfolio value.
34. Financed tonnes and carbon footprint
Financed emissions attribute part of an issuer's emissions to an investment using an attribution factor. In a simplified single-holding problem, divide invested value by the specified issuer valuation denominator, then multiply by issuer emissions. Sum attributed tonnes across holdings for total financed emissions; divide by portfolio value in millions for a carbon footprint per $1 million invested. State the denominator convention, valuation date, emissions scope, and data coverage, since these choices affect results. Unlike WACI, this method yields attributed tonnes before normalization. Market-value changes can move the attribution factor even when the issuer's real emissions are unchanged. Do not read a lower footprint as proof of decarbonization without checking both numerator and denominator. Attributed tonnes are an accounting measure, not physically separable emissions.
Worked example: A hypothetical investor holds $6 million in an issuer valued at $60 million under the problem's attribution convention. The issuer emits 100,000 tCO2e. Attribution is $6 million ÷ $60 million = 10%, giving 10,000 tCO2e financed emissions. For this single holding, footprint is 10,000 ÷ 6 = about 1,667 tCO2e per $1 million invested.
Watch the distinction: The 10,000 tonnes are attributed emissions, while 1,667 is a normalized footprint. Neither equals the issuer's emissions intensity per dollar of revenue.
35. Choosing and checking carbon metrics
Absolute emissions measure total tonnes within a stated boundary. Intensity divides tonnes by an activity measure such as revenue or output. The pair answers different questions: absolute emissions describe contribution to atmospheric loading, while intensity can show efficiency relative to economic activity. A falling intensity can accompany flat or rising absolute emissions if the denominator grows faster. Conversely, revenue intensity can rise during a price downturn even with stable physical output and emissions. Choose a denominator that fits the decision and keep units visible. Also inspect which scopes, issuers, and years are covered. Missing high-emission data or replacing measured figures with estimates can change a portfolio metric. When comparing periods, separate genuine operating change from shifts in coverage, methods, prices, and holdings.
Worked example: An issuer emits 100 tCO2e in each of two hypothetical years. Revenue rises from $1 million to $2 million. Revenue intensity falls from 100 ÷ 1 = 100 to 100 ÷ 2 = 50 tCO2e per $1 million, a 50% decline. Absolute emissions remain 100 tCO2e, so the lower intensity alone shows no reduction in total emissions.
Watch the distinction: A lower carbon intensity is not automatically an absolute emissions cut. Inspect the emissions numerator and the activity denominator separately.
36. Risk appetite and monitoring
Risk appetite becomes useful when it is translated into a measurable exposure definition, a limit, an owner, and a response when the limit is crossed. Define what counts as highly exposed collateral, the portfolio denominator, how often both are refreshed, and whether missing location data are treated separately. A breach should trigger investigation and an authorized decision: reduce exposure, improve protection, change lending terms, or approve a documented exception within governance rules. Monitoring then checks whether the action actually lowers risk. A limit is an internal management choice, not a universal safety threshold. It may need review as hazards, asset values, or portfolio composition change. Pair concentration measures with severity and repayment analysis, since the same percentage can represent very different expected losses.
Worked example: A firm's hypothetical limit is 15% of collateral value in its defined high-flood-exposure category. Current measurement is 18%, an excess of 18% − 15% = 3 percentage points. Management identifies which loans drive the excess, tests their defenses and maturity dates, and proposes a reduction or documented exception. The next monitoring cycle checks both the percentage and the agreed action.
Watch the distinction: A breached limit is a decision trigger, not proof that every exposed loan will fail. Ignoring the breach because the current loss is zero defeats monitoring.
Reference: [1]
Climate Models and Scenario Analysis
37. What climate models represent
Physical climate models represent processes such as radiation, circulation, and the water cycle to estimate climate conditions under specified inputs. Integrated assessment models connect economic activity, energy systems, emissions, and climate pathways, often producing variables such as energy mix, emissions, and carbon prices. For financial analysis, follow the handoff: choose a coherent pathway, obtain relevant climate or economic variables, translate them to local hazards or sector costs, and map those changes to an asset's cash flows. Each step adds assumptions. A global temperature result does not directly give flood depth at a factory, and a carbon-price path does not directly give a borrower's profit margin. Models support conditional exploration of possible futures, so check spatial scale, time horizon, and variable definitions before combining outputs.
Worked example: An integrated assessment model gives a hypothetical emissions path and future energy prices. A physical model uses the emissions pathway to estimate regional rainfall changes. To assess one factory, an analyst then needs local terrain, drainage, flood defenses, and damage functions. Global GDP and temperature outputs alone cannot determine that factory's flood depth or resulting loan loss.
Watch the distinction: Treating every climate-model output as a site-level forecast skips essential translation. Scale and asset vulnerability often determine whether the output is usable for a decision.
Reference: [1]
38. Downscaling and local exposure
Downscaling takes broad climate-model information toward a finer spatial scale, using statistical relationships or higher-resolution physical modeling. It can improve relevance for local decisions but cannot create certainty that the parent model or local data lack. For a flood decision, connect regional rainfall or sea-level information with catchment behavior, drainage, terrain elevation, defenses, and the asset's floor level. Check whether elevation and hazard data share a vertical reference and whether defenses are maintained; small differences can materially alter estimated inundation. Separate hazard from exposure: two assets in one model grid cell can have different damage because they occupy different ground levels or use different protection. Test multiple plausible inputs where local measurements or future conditions are uncertain.
Worked example: Two hypothetical facilities lie in the same coarse climate-model cell. One floor is 1 meter above the chosen local reference level; the other is 3 meters above it. If modeled flood water reaches 2 meters on that same reference, the first faces 1 meter of water above its floor while the second remains dry, assuming no other flow path or defenses.
Watch the distinction: A finer map is not automatically a more accurate forecast. Its local terrain, drainage, reference levels, and underlying climate assumptions still need checking.
Reference: [1]
39. Orderly, disorderly, and hot-house scenarios
Scenario families organize different combinations of transition and physical pressure. An orderly pathway assumes earlier, more coordinated action, so policy and technology changes can be spread over time while longer-run warming is restrained. A disorderly pathway introduces delayed or uneven action, which can produce sharper transition shocks when adjustment accelerates. A hot-house pathway assumes weak mitigation and therefore greater long-run physical pressure, even if near-term transition costs appear smaller. These are conditional narratives with modeled variables, not assigned probabilities or promises about the future. Compare them against the same asset, geography, and horizon. A short loan may be sensitive to near-term carbon costs; a long-lived coastal asset may be more sensitive to later sea-level and extreme-event exposure. The SCR study materials list scenario analysis among the curriculum chapters.
Worked example: A power company has a loan refinancing in 2032 and a coastal plant operating beyond 2040. A hypothetical late policy tightening in 2030 may weaken cash flow before refinancing, making the disorderly pathway important for credit. Weak mitigation may create greater later flood exposure for the plant. Neither pathway's label says which will occur; each reveals a different vulnerability.
Watch the distinction: Do not rank scenarios as most or least likely from their names. Use their timing and assumptions to test decisions under different pressures.
Reference: [1]
40. Translating scenarios into value
To turn a scenario into value, translate its drivers into incremental cash flows, then discount each change at the end of its period under a stated rate. For example, a carbon price may raise costs, while pass-through or lower energy use may offset some of the effect. Use the net cash-flow difference from a consistent baseline, not the scenario's headline variable. The present-value effect equals the sum of each year's cash-flow change divided by one plus the discount rate raised to that year's power. Check whether effects persist beyond the explicit forecast, whether capital spending or taxes change, and whether the chosen discount rate already reflects risks being modeled in cash flows. Scenario values are conditional estimates; sensitivity to price, timing, and discount assumptions often matters more than a single result.
Worked example: A hypothetical shock reduces end-of-year cash flow by $4 million in each of the next three years, with no later effect. At a 10% annual discount rate, present-value loss is $4 million × (1/1.10 + 1/1.10² + 1/1.10³) = $4 million × (0.9091 + 0.8264 + 0.7513) ≈ $9.95 million. That is the modeled value decrease under these assumptions.
Watch the distinction: Adding three $4 million losses gives $12 million undiscounted. For today's value, discount each year's loss according to when it occurs.
Reference: [1]
41. Timing and financial horizons
A scenario's endpoint can conceal the years when losses actually arise. For a lender, map each policy or hazard change onto the loan's payment dates, maturity and likely refinancing date. Then ask whether the borrower can meet obligations during the shock, even if its business is viable later. For an investor, compare that timing with the asset's useful life and expected sale date. A long lived plant may face costs after a short loan matures, yet those costs can still affect today's collateral value and the borrower's ability to refinance. Two pathways that reach the same emissions level can therefore produce different credit outcomes. The SCR study materials place scenario analysis within the current curriculum; the practical task is to examine the path between dates, not just its final value.
Worked example: Two hypothetical pathways both reach 20,000 tonnes of annual emissions in 2040. Path A brings a sharp policy cost increase in 2030; Path B phases it in after 2035. A loan matures in 2032 and needs refinancing. Under A, the lender must test whether the borrower can absorb higher costs before refinancing. Under B, that immediate refinancing pressure is smaller, despite the identical 2040 endpoint.
Watch the distinction: A matching 2040 endpoint does not imply matching risk. Cash shortages, covenant breaches and refinancing decisions can occur years before the endpoint.
Reference: [1]
42. Scenario uncertainty and sensitivity
Scenario results depend on assumptions about hazards, local defenses, asset vulnerability, adaptation performance and economic conditions. Sensitivity analysis changes one or more assumptions to show which ones drive a decision. Start by specifying the loss measure: damage from a selected event, expected annual loss or a discounted lifetime amount. Compare like with like, then test whether an adaptation remains useful across plausible cases. A wide range signals uncertainty, not a probability distribution unless probabilities have been justified. It is especially useful to identify a decision threshold: the assumptions under which the preferred action changes. This reasoning resembles the treatment of uncertain losses in the Associate in Risk Management guide. Communicate residual exposure after any intervention so a lower modeled loss is not mistaken for safety under every possible event.
Worked example: A hypothetical site study estimates damage from a specified flood event between $2 million and $20 million. A $5 million barrier lowers the upper event loss to $8 million, a $12 million reduction in that upper case. The lower case alone cannot justify the barrier, and the $12 million is not an expected saving. Managers need event likelihood, performance across cases and asset life before comparing benefits with cost.
Watch the distinction: Do not subtract a construction cost from the highest modeled loss and call the result net value. One event estimate is not an expected lifetime benefit.
Reference: [1]
Net Zero
43. Net-zero target boundaries
A net-zero target has meaning only after you identify its accounting boundary. Scope 1 covers direct emissions from sources the organization owns or controls; Scope 2 covers emissions from purchased energy; Scope 3 covers other value-chain emissions, such as purchased goods and use of sold products. Check which entities, activities, gases and scopes the target includes, and whether exclusions are disclosed. The base year and measurement method also matter because an acquisition or boundary change can alter reported totals without changing underlying performance. When Scope 3 dominates, an operational target can cover only a small share of the organization's climate footprint. Evaluate both the target's covered share and the practical influence the organization has over those emissions. Coverage alone says little about the speed or credibility of planned reductions.
Worked example: A hypothetical company reports 4,000 tCO2e in Scope 1, 6,000 tCO2e in Scope 2 and 90,000 tCO2e in Scope 3. Total reported emissions are 100,000 tCO2e. A target covering only Scopes 1 and 2 covers 10,000 ÷ 100,000 = 10% of that total. The target may guide plant and electricity decisions, but it leaves 90% outside its stated boundary.
Watch the distinction: A target described as company-wide may still exclude most reported emissions. Read the scope and entity boundaries before judging its reach.
Reference: [1]
44. Absolute and intensity targets
An absolute target tracks total emissions, usually in tonnes of CO2 equivalent. An intensity target divides emissions by a measure of activity, such as units produced, revenue or energy generated. Intensity helps compare operations at different scales, but its denominator can change independently of emissions. A growing business can lower emissions per unit while its total emissions rise; shrinking output can lower total emissions while making intensity worse. Calculate both measures against a consistent base year, using the same operational boundary and units. For intensity, state exactly what the denominator represents, since revenue can move with prices and exchange rates. When judging progress toward a climate outcome, read intensity alongside absolute emissions and expected production. A credible interpretation explains what changed in technology or operations as well as what changed in scale.
Worked example: A hypothetical plant emits 100,000 tonnes while producing 1 million units, then emits 90,000 tonnes while producing 1.5 million units. Absolute emissions fall by (100,000 − 90,000) ÷ 100,000 = 10%. Intensity falls from 0.10 to 0.06 tonne per unit, a (0.10 − 0.06) ÷ 0.10 = 40% reduction. Production growth explains why the two percentages differ.
Watch the distinction: A 40% intensity improvement is not a 40% cut in total emissions. The production denominator grew, while total emissions fell by only 10%.
Reference: [1]
45. Cumulative emissions pathways
CO2 has a cumulative effect: emissions in earlier years continue adding to the atmospheric stock. For that reason, a pathway's total emissions before its end date can matter more than the final year's rate alone. To compare paths, put annual emissions on the same boundary, add the tonnes for each year and check whether interim changes are plausible. The simple sum is a discrete approximation to the area under an emissions curve; it is not an estimate of global temperature by itself. Earlier reductions lower cumulative emissions even when two paths finish at the same annual level. This also makes interim milestones useful: they reveal whether a distant commitment is being delayed. When interpreting a pathway, separate gross emissions cuts from credits or removals, and disclose assumptions about production growth, asset retirement and technologies needed later.
Worked example: Two hypothetical five-year paths both end at 2,000 tonnes annually. Path A emits 10,000, 10,000, 10,000, 10,000 and 2,000 tonnes, totaling 42,000 tonnes. Path B emits 10,000, 8,000, 6,000, 4,000 and 2,000 tonnes, totaling 30,000 tonnes. With identical accounting boundaries, B emits 12,000 fewer tonnes across those five years because its cuts occur earlier.
Watch the distinction: Matching final-year emissions do not mean equal climate effect. Earlier emissions add to the cumulative total and cannot be erased by a late annual cut.
Reference: [1]
46. Reductions, credits, and removals
A reduction lowers emissions from an activity against its own previous or projected level. An avoided-emissions credit usually represents a claimed reduction relative to a counterfactual elsewhere. A removal takes greenhouse gas out of the atmosphere and stores it. These are different claims and should be recorded separately from the purchaser's gross inventory. For a credit, examine the counterfactual, whether the activity was additional, measurement quality, double counting and the risk that the benefit reverses. For a removal, also examine how long storage lasts and who bears reversal risk. A forest project and geological storage may have different durability and monitoring needs. The label on a purchased unit does not establish its quality. When assessing a net-zero claim, first calculate gross emissions and actual internal cuts, then identify the specific residual emissions and evidence for any units used against them.
Worked example: A hypothetical firm emits 1,000 tCO2e from its operations and buys a credit for 1,000 tCO2e allegedly avoided by a separate project. Its gross operational inventory remains 1,000 tCO2e. If it instead claims a removal, it must show that atmospheric carbon was captured, the quantity was measured and storage is durable. Neither purchase demonstrates that its equipment emitted less.
Watch the distinction: A credit purchase is not an operating reduction. Net claims can include separate instruments, but gross emissions must remain visible for readers to assess actual cuts.
Reference: [1]
47. Portfolio net-zero accounting
Financed emissions attribute part of an investee's emissions to a financial institution. In a simple equity example, attributed tonnes equal the holding's value divided by the issuer's enterprise value, multiplied by issuer emissions. This measures a share under a stated accounting method; it does not mean the investor physically caused that exact amount. The numerator can change through buying or selling, while the denominator can change through valuation and financing, even if the issuer's factories emit the same amount. Therefore, decompose year-to-year movement into portfolio activity, attribution factors, issuer emissions and data revisions. Portfolio targets should be judged alongside evidence of real-world changes at issuers, such as lower production emissions and funded transition actions. The financial risk manager guide offers related financial-risk context, while this calculation needs its own boundary and valuation assumptions.
Worked example: A hypothetical $10 million holding represents 10% of an issuer valued at $100 million. With issuer emissions of 50,000 tCO2e, attributed emissions are 0.10 × 50,000 = 5,000 tCO2e. If enterprise value rises to $200 million while the holding stays $10 million, attribution becomes 5% × 50,000 = 2,500 tCO2e. The issuer still emits 50,000 tCO2e.
Watch the distinction: A fall in attributed portfolio tonnes does not prove the investee cut emissions. First check valuation, holdings and data changes.
Reference: [1]
48. Testing net-zero credibility
Test a net-zero commitment by linking its endpoint to a measured baseline, interim milestones and actions that can deliver them. Identify which emissions sources each action affects, its expected tonnes saved, funding, implementation date and responsible owner. Then compare the sum of supported cuts with the required reduction, allowing for overlap between measures. Examine dependencies such as grid changes, supplier participation or an unproven technology; a plan can be exposed even when its arithmetic balances. Residual emissions require a separate explanation of why they remain and how any claimed removals will be measured and maintained. Progress reports should reconcile changes in the inventory boundary and data quality so apparent improvement is not merely reclassification. A target becomes more credible when the near-term investment and operational decisions are consistent with its emissions pathway.
Worked example: A hypothetical firm emits 100,000 tCO2e and targets 60,000, requiring a 40,000-tonne cut. Funded equipment and energy measures support 10,000 tonnes, leaving 30,000 tonnes unsupported. If a proposed supplier program could cut 15,000 tonnes but has no contracts or milestones, it is still a dependency, not a delivered reduction. The target's arithmetic is clear; its implementation evidence is incomplete.
Watch the distinction: A precise target year and percentage do not prove feasibility. Check funded measures, timing, dependencies and treatment of remaining emissions.
Reference: [1]
Climate and Nature Risk Assessment
49. Nature dependencies and impacts
An organization depends on nature when its operations need ecosystem services such as water supply, pollination, soil quality or flood protection. It affects nature through pressures such as land conversion, extraction, pollution and habitat fragmentation. The two directions can reinforce each other: wastewater can degrade the river a factory needs, while water scarcity can restrict its output. Begin with a specific activity and location, identify the relevant ecosystem service, then trace how a change in that service affects production, costs or assets. Separately identify the organization's pressure and the resulting ecological condition. This helps distinguish a business risk from an impact on people and ecosystems, while recognizing that an impact can later create financial risk. The ESG investing guide provides related investment context; location and ecological conditions remain essential for nature assessment.
Worked example: A hypothetical beverage plant draws water from a river and releases treated wastewater into it. A drought reduces river flow, limiting production: that is a dependence becoming a business risk. If the plant's discharge worsens water quality downstream, that is an impact. The impact may later intensify its own water-treatment costs, so both directions belong in the assessment.
Watch the distinction: Calling every nature issue an external threat misses the organization's own pressures. Record both what the business needs from nature and what it changes.
Reference: [1]
50. Pressures versus ecosystem condition
A pressure measure describes an activity that may affect nature, such as water withdrawal, nutrient discharge or land clearing. An ecosystem-condition measure describes what is happening to the system, such as species composition, habitat quality or water quality. A response measure records what was done, such as planting trees or restoring a stream. These measures answer different questions. Reducing a pressure can improve conditions, but ecological recovery may take time, face other pressures or fail if the intervention is poorly suited to the site. Use a causal chain: identify the pressure, establish a relevant baseline condition, describe the action and monitor whether condition changes. Choose indicators that reflect the actual ecosystem and location. An area planted is easy to count, yet it may conceal low survival, invasive species or loss of native habitat.
Worked example: A hypothetical company plants 100 hectares of trees and reports 100 hectares treated. That number measures its intervention. If 60 hectares survive but monitoring finds few native species, the survival area and biodiversity condition tell a more qualified story. To claim ecological recovery, the company would need evidence that relevant habitat quality or species measures improved against a suitable baseline.
Watch the distinction: An intervention count is not an outcome. Planted hectares, money spent or suppliers contacted cannot by themselves establish biodiversity recovery.
Reference: [1]
51. The TNFD LEAP sequence
The TNFD LEAP approach provides a sequence for investigating nature-related issues. Locate identifies where an organization's activities and value chain meet ecosystems, especially sensitive or degraded places. Evaluate examines dependencies on ecosystem services and impacts on nature at those locations. Assess translates material dependencies and impacts into risks and opportunities for the organization, including possible effects on operations and finance. Prepare turns findings into management responses, governance, metrics and disclosures. The sequence matters because nature risk is often highly local: an average sector score can miss a supplier drawing water from a stressed basin. Iterate when new supplier or site data changes the map. LEAP supports structured judgment; it does not supply a universal numeric score or eliminate uncertainty about ecological change.
Worked example: A hypothetical food manufacturer maps farms supplying its main ingredient and finds several in a stressed watershed. It evaluates reliance on irrigation and effects from withdrawals, then assesses possible crop shortfalls and sourcing costs. It prepares supplier water actions, monitoring and disclosure of material findings. Starting with aggregate company emissions alone would not reveal which farms face this local water risk.
Watch the distinction: LEAP is not a checklist that turns location data directly into a financial loss. Dependencies and impacts must be evaluated before risk is assessed.
Reference: [1]
52. Mapping nature-sensitive locations
Nature exposure is tied to place. Map owned sites and, where possible, important suppliers against watersheds, habitat condition and ecologically sensitive areas. Then add business significance: purchase share, substitutability, inventory buffers and the time needed to switch supply. Proximity to a sensitive area is a screening cue, not proof of harm or loss. Site activities determine the pressure, while local ecological condition affects both impact and dependence. Missing supplier coordinates should be flagged because a country-level average can hide a concentrated exposure. A useful prioritization asks three questions: Is the location ecologically sensitive? Does the activity depend on or pressure that ecosystem? Would disruption materially affect the organization? Investigate high-priority combinations before assigning detailed financial values. Revisit the map as sourcing changes and as better location data becomes available.
Worked example: A hypothetical manufacturer buys 20% of a critical input from a supplier beside a protected wetland. Another supplier of equal purchase value operates away from identified sensitive habitat. The first site merits closer review of water use, discharge and sourcing alternatives. The 20% share shows business significance, while the location is a screening signal; neither alone proves the supplier is damaging the wetland.
Watch the distinction: A map pin near a protected area is not a quantified loss. Check the activity, ecological pathway and business dependence before estimating financial exposure.
Reference: [1]
53. Nature risk in financial statements
To connect nature deterioration with creditor risk, trace a physical or ecological change through operations into cash flow. Pollinator decline, for example, may lower crop yield, increase replacement costs or force changes in land management. Estimate the resulting revenue and cost effects, then recalculate earnings, interest coverage and cash available for debt service. Check whether the borrower can adapt, pass on costs or use reserves; these responses may reduce or delay loss. A ratio change is an indicator, not a default probability. Lenders also need debt maturity, collateral value, covenants and the persistence of the ecological problem. Avoid treating every observed yield change as caused by nature loss, since weather, prices and farm practices may also contribute. The analysis is strongest when the causal link and assumptions are explicit.
Worked example: A hypothetical grower has $10 million annual EBITDA and $4 million annual interest, so interest coverage is 10 ÷ 4 = 2.5 times. If pollinator loss reduces output and EBITDA to $7 million, coverage becomes 7 ÷ 4 = 1.75 times. The borrower has less earnings headroom for interest, though this ratio alone does not establish whether it will default.
Watch the distinction: A lower coverage ratio is evidence of weaker debt capacity, not a computed default rate. Test other cash needs, adaptation options and causal assumptions.
Reference: [1]
54. Interacting climate and nature losses
Climate and nature risks can share a causal pathway. Wetland loss may increase flood depth during a storm, while hotter conditions can worsen drought and ecosystem decline. When two models estimate losses from the same affected buildings, crops or period, summing their outputs may count damage twice. First define the event, asset boundary, time horizon and loss components in each estimate. Then identify which losses overlap and which are genuinely additional. Where possible, model a common baseline, a climate-only case and a combined climate-and-nature case. The increment between relevant cases shows the additional effect under those assumptions. Interactions can also amplify harm, so simply choosing the larger standalone estimate may understate combined risk. Explain uncertainty in the interaction and whether proposed adaptation addresses the shared cause.
Worked example: A hypothetical flood model estimates $8 million of building damage. A separate wetland-loss study reports $5 million of flood-related damage to the same buildings under the same storm. Adding them gives $13 million, but the studies may overlap. If a combined model estimates $11 million under aligned assumptions, the incremental loss relative to the $8 million baseline is $3 million, not $5 million.
Watch the distinction: Separate report totals are not automatically additive. Align the event and asset boundaries, then measure any incremental damage from the interaction.
Reference: [1]
Transition Planning and Carbon Reporting
55. Transition-plan structure
An executable transition plan connects an emissions baseline to targets, actions and money. Start with a consistent inventory and identify major emissions sources. Set interim milestones, then assign operational levers such as efficiency, equipment replacement, purchased-energy changes or supplier engagement to those sources. Estimate each lever's emissions effect, cost, timing and dependencies, taking care not to count the same reduction twice. Match capital spending to normal replacement cycles where feasible, while checking whether waiting would make milestones unattainable. Name owners who can approve investment and monitor delivery, and define what triggers revision if performance differs from the plan. Reporting should distinguish completed actions from planned ones and reconcile observed emissions with expected cuts. A plan can be ambitious and still fail if its investments arrive after the emissions milestone.
Worked example: A hypothetical manufacturer emits 50,000 tCO2e annually and targets 35,000 by 2030, a 15,000-tonne cut. It schedules a furnace replacement expected to save 9,000 tonnes and electricity changes expected to save 6,000, totaling 15,000 if effects do not overlap. Funding, installation dates and milestone reviews must precede 2030. A target with the same arithmetic but no approved furnace investment lacks an execution path.
Watch the distinction: A target is not a plan. Check that specific actions, investment dates and accountable owners can deliver the stated cuts on time.
Reference: [1]
56. Prioritizing transition investment
A simple abatement-cost measure divides a project's incremental cost by the emissions it is expected to avoid over a defined period. Keep the numerator and denominator consistent: use lifetime costs with lifetime tonnes, and account for operating savings, maintenance, timing and discounting when the decision warrants them. Cost per tonne is useful for ranking, but a lower figure does not settle the choice. Projects can differ in technical readiness, dependence on grid changes, implementation time, operational resilience and ability to meet an interim milestone. A capital budget may also prevent selecting an otherwise attractive project. Compare feasible options against the same baseline and avoid double counting savings when projects interact. After selecting a project, track actual expenditure and measured emissions to see whether the expected cost and abatement were delivered.
Worked example: Hypothetical Project A costs $2 million net and avoids 50,000 lifetime tCO2e: $2 million ÷ 50,000 = $40 per tonne. Project B costs $3 million net and avoids 30,000 tonnes: $100 per tonne. With a $2 million capital limit, A fits and has the lower simple cost. The ranking assumes comparable baselines and credible lifetime savings; feasibility and timing still need review.
Watch the distinction: The lowest quoted dollars per tonne can be misleading if projects use different lifetimes, omit operating costs or cannot be built before the target date.
Reference: [1]
57. People affected by transition
Transition decisions affect people as well as emissions. A closure or technology change may alter employment, customer prices, service access and a community's tax base. Map who is affected, when effects occur and what practical options exist for each group. For workers, distinguish retraining with plausible jobs from an unallocated promise to provide support. For customers, examine affordability and whether essential service remains available. For communities, consider local dependence on a site and whether alternative economic activity is credible. These factors can change implementation risk through delays, opposition, workforce shortages or loss of trust. They also matter as outcomes in their own right. A useful plan assigns resources, dates and owners to the measures, then monitors whether affected groups actually receive the intended support.
Worked example: A hypothetical plant closure affects 400 workers. A funded retraining program has places for 250 and confirmed redeployment covers 100 more. That leaves 400 − 250 − 100 = 50 workers without a stated pathway, assuming the groups do not overlap. Management should address those 50 and check whether training leads to viable work, even if the closure achieves its emissions target.
Watch the distinction: Meeting an emissions target does not show that transition effects on people were managed. Count uncovered groups and test whether proposed support is usable.
Reference: [1]
58. Purchased-electricity emissions
Purchased-electricity emissions are calculated from electricity consumed multiplied by an emissions factor, with matching units and reporting period. A location-based figure uses a factor representing the relevant grid's generation mix. A market-based figure uses qualifying contractual information where available and applicable; its validity depends on the accounting method's quality criteria and the contracts' attributes. The two figures can differ because they answer different questions about purchased energy. Record meter data, factor source, geography, period and any contractual evidence. Avoid treating a low contractual factor as proof that the local grid physically produced no emissions at the time of use. If renewable purchases also change operations or total electricity consumption, separate those effects in the reconciliation. The SCR required readings include greenhouse-gas accounting principles and boundaries relevant to evaluating such inventories.
Worked example: A hypothetical site consumes 1,000 MWh. Using a grid factor of 0.4 tCO2e/MWh gives 1,000 × 0.4 = 400 tCO2e on a location basis. If qualifying contractual evidence supports 0.1 tCO2e/MWh for the same 1,000 MWh, the market-based figure is 100 tCO2e. Report both with their methods; the 300-tonne difference is not a fall in metered consumption.
Watch the distinction: A contractual Scope 2 figure and a grid-based figure are not interchangeable. Label the method and retain the underlying MWh and evidence.
Reference: [1]
59. Scope 3 estimates and revisions
Scope 3 estimates often begin with activity data, such as quantities purchased, multiplied by emissions factors representing upstream or downstream processes. A generic factor can support an initial estimate, but it may hide differences among suppliers, production methods and locations. Better supplier data can improve specificity if its boundary, units, period and verification are sound. When a factor changes, explain whether the revised tonnes reflect real operational improvement, a better measurement method or both. For comparisons over time, consider recalculating the base year under a consistent method where appropriate; otherwise a data improvement may appear to be an emissions cut. Keep a clear audit trail from purchase records through factors to totals, including assumptions for missing categories. Prioritize data improvements where emissions are large and the estimate is uncertain.
Worked example: A hypothetical buyer purchases 2,000 units. A generic factor of 2 tCO2e per unit gives 4,000 tCO2e. Verified supplier data supports 1.5 tCO2e per unit for the same goods and boundary, giving 3,000 tCO2e. The estimate falls by 1,000 tCO2e, or 25%. With quantity unchanged, this revision alone does not prove the supplier's production improved during the year.
Watch the distinction: A lower estimate after changing factors is not automatically decarbonization. Separate methodological revision from an evidenced change in the supplier's actual process.
Reference: [1]
60. Reporting claims and assurance evidence
A defensible public emissions claim reconciles several records without collapsing them into one number. The gross inventory records emissions within a stated organizational and operational boundary. A target states a future or interim outcome against a defined baseline. Credits and removals are separate instruments that may be used in a net claim, subject to their own quality and retirement evidence. For assurance, retain activity data, factors, calculation methods, boundary decisions, source documents and approval records. Reconcile current gross emissions with prior years, explaining structural or methodological changes. Then show credits separately, including quantities, vintage, project information and retirement status where relevant. Check that public wording matches what the evidence supports: an inventory decrease, a reduction against a target and a credit-backed claim are distinct. Clear separation lets readers assess actual operational progress and the role of external instruments.
Worked example: A hypothetical firm reports 20,000 tCO2e of gross emissions and retires credits representing 5,000 tCO2e. Its gross inventory remains 20,000 tCO2e. If it presents an adjusted claim of 15,000 tCO2e, it must label that arithmetic and explain the credits separately. Calling 15,000 tCO2e its gross emissions would misstate the inventory and obscure whether its operations improved.
Watch the distinction: Subtracting retired credits from gross inventory changes the meaning of the number. Preserve gross tonnes, disclose instruments separately and support each claim with records.
Reference: [1]
Official references and further reading
- SCR® Study Material, Books, Digital Learning Platform | GARP
- SCR® Exam Information, Steps to Earn Certificate | GARP
- Required Readings | Sustainability and Climate Risk (SCR) | GARP