A research department produces opinions on securities and markets and sells them, directly or indirectly, to investors. In a corporate and investment bank it sits in Global Markets, next to sales and trading, and not in investment banking. The separation is deliberate. Research earns its credibility from investors only if they believe the analyst would publish a Sell on a company that the bank’s bankers are trying to win as a client, so the rules place research on the public side of the information barrier, keep it out of pitches and protect its budget from deal revenues.
The teams. Four families coexist. Equity research is organised by sector: a metals and mining team, a utilities team, a banks team, and so on, each with a lead analyst, one or two associates and a junior. Credit research covers the bonds and loans of issuers, again by sector, and splits into two kinds of analyst. Independent credit research publishes recommendations under the same rules as equity research. Desk analysts (also called desk strategists) sit with the credit traders and write short commentaries for them and their clients; their output is usually classified as a marketing communication rather than investment research, which matters for the rules that apply. Macro, rates and FX strategy covers economies, central banks and curves. Finally, most large brokers now have an ESG and climate research team: a handful of specialists who write thematic and policy reports (the EU carbon market, CBAM, the taxonomy, transition plans), maintain the house ESG and carbon datasets, and support sector analysts when an ESG question becomes material to a stock or a bond.
Clients. Research is read by institutional investors: long-only asset managers, hedge funds, insurers and pension funds, and the investment teams of private banks. Within those firms the readers are portfolio managers, buy-side analysts, and increasingly stewardship and ESG teams who prepare votes and engagement. Internal clients matter too: sales people use research to call clients, traders use it to understand flows, and the bank’s own credit risk function reads it, without being allowed to steer it.
How research earns money. Research rarely books revenue like a trading desk. Since MiFID II, an EU investment firm that receives research must pay for it either from its own resources or from a research payment account funded by its clients, so many large asset managers sign annual research contracts with brokers, for a fee that ranges from tens of thousands to a few million euros depending on the broker and the depth of service. The rest of the value is indirect: research generates trading commissions in cash equities and flows in credit, and it builds the franchise that wins corporate access events and investor votes. Costs are almost entirely people and data: an experienced lead analyst, associates, licences for financial and ESG data, and a share of publishing and compliance infrastructure. For most banks, research is a cost centre whose value is argued for through attribution, not measured by a profit and loss account.
Adjacent functions. Research compliance (sometimes called supervisory analysts or publishing) reviews every report before release: disclosures, conflicts, consistency of ratings and target prices, and the absence of inside information. The control room keeps the watch list and restricted list and decides whether an analyst can be “wall-crossed” for a confidential transaction. Equity and credit sales distribute the research and gather feedback. Corporate access organises roadshows and conferences. The data team runs the models, the consensus feeds and the vendor licences.
A realistic week. Monday, 6:45: the morning meeting. Each analyst with news has two minutes to tell sales what changed and what to say to clients; the metals and mining analyst explains that the EU carbon price rose overnight and what it means for steel margins. Tuesday: a company reports quarterly results; the analyst updates the earnings model, publishes a results note within hours, then joins calls with portfolio managers. Wednesday: the ESG research team publishes a policy note on the Council’s negotiating position on the EU ETS review, and the steel analyst is asked to quantify the effect on the companies covered. Thursday: the analyst spends the day with a client in Frankfurt and holds four one-hour meetings, the “marketing” of a recent report. Friday: research compliance returns the draft of an initiation of coverage with twenty comments; the analyst also prepares the annual broker review, where clients vote on the usefulness of each broker’s research and set next year’s payments.
WHO DOES WHAT — The lead sector analyst owns the rating, the target price and the model. The associate builds and maintains the model and drafts. The ESG research specialist supplies carbon and policy inputs and reviews the ESG section, but does not set the rating. Research compliance approves publication. The head of research arbitrates disputes on coverage and resources, never on the content of a recommendation in response to a banker.
The flagship product of an equity analyst is the initiation of coverage: a 40 to 80 page report that presents the investment case, the industry, the company, a full earnings model, a valuation, a rating (Buy, Hold, Sell, or Overweight, Equal-weight, Underweight) and a 12-month target price. Afterwards, the analyst publishes company updates (results previews and reviews, reactions to news), sector reports comparing companies, and thematic reports on a cross-cutting question. Each report carries a disclosure appendix that is often longer than the summary.
The credit analyst’s equivalent is a credit initiation or issuer profile: business risk, financial risk, liquidity, structure of the debt (seniority, guarantees, covenants), a view on the issuer’s rating trajectory, and recommendations, generally relative: overweight or underweight the issuer’s bonds against its sector or an index, or a switch between two points on its curve. Credit research cares less about upside than about the probability and severity of bad outcomes, which is why the two analysts can read the same company differently, as the case will show.
Most sector analysts use two methods and reconcile them. The first is a multiple: enterprise value is a multiple of a cash-earnings measure, typically EBITDA for capital-intensive industries such as steel, chosen by reference to the company’s own history and its peers. The second is a discounted cash flow (DCF) model: forecast free cash flow to the firm for several years, add a terminal value and discount at the weighted average cost of capital.
IN ONE SENTENCE — A target price is the equity value per share that the analyst expects the market to recognise within twelve months, obtained by valuing the enterprise and subtracting everything that ranks ahead of shareholders.
INTUITION — Valuing equity is like valuing a house with a mortgage. You estimate what the house is worth (enterprise value), subtract what is owed to the bank (net debt and debt-like items), and divide the rest among the owners. A small change in the house price produces a larger percentage change in the owners’ equity when the mortgage is large.
FORMULA
is the target price per share; the enterprise value; net financial debt; debt-like items such as pension deficits and provisions; the number of shares; the EV/EBITDA multiple; free cash flow to the firm in year (EBITDA minus tax on operating profit minus capital expenditure minus change in working capital); the weighted average cost of capital; the perpetual growth rate after year .
The rating then follows from the expected return to the target price. Each broker publishes its own rating system in its disclosures; Meridian Research, in this course, rates a stock Buy if the upside to the target price exceeds +15%, Sell if it is below −15%, and Hold in between.
INTERACTIVE LAB — From EBITDA to a target price and a rating
See how the multiple, the debt and the share price determine the target price, the upside and the rating, and why equity is geared to enterprise value.
| Parameter you can move | Range | Default |
|---|---|---|
| EBITDA used for valuation | 100 to 1500 EUR m | 500 EUR m |
| Target EV/EBITDA multiple | 3 to 12 x | 6 x |
| Net debt | 0 to 3000 EUR m | 1000 EUR m |
| Debt-like items (pensions, provisions) | 0 to 1000 EUR m | 200 EUR m |
| Shares outstanding | 10 to 1000 m | 100 m |
| Current share price | 1 to 100 EUR | 15 EUR |
| Result at default settings | Value |
|---|---|
| Target enterprise value | 3,000 EUR m |
| Target equity value | 1,800 EUR m |
| Target price | 18.00 EUR |
| Upside to target price | 20.0 % |
| Rating (1 Buy, 0 Hold, -1 Sell) | 1 |
| EV/EBITDA implied by the current price | 5.40 x |
| Change in target price for a 1% change in EBITDA | 1.67 % |
Try it. Cut EBITDA by 10% to 450: the target price falls by 16.7%, to exactly the current price, and the Buy becomes a Hold. That leverage is why a carbon cost that looks small against EBITDA can be large against equity.
PITFALL — Students often apply the multiple to the wrong earnings. A mid-cycle multiple applied to peak-cycle EBITDA, or a historical multiple applied to an EBITDA that will soon bear a new cost, gives a target price that is consistent in form and wrong in substance. For a steelmaker in 2026, the question “which EBITDA?” is precisely where carbon enters.
A credit analyst asks whether the issuer will pay interest and principal on time, and what bondholders recover if it does not. The work has three blocks. Business risk: industry cyclicality, competitive position, cost position on the industry cost curve, diversification. Financial risk: leverage and coverage ratios, typically adjusted debt over EBITDA (adjusted debt adds pension deficits and leases to net debt), funds from operations (FFO, roughly EBITDA minus interest and cash taxes) over adjusted debt, and EBITDA over interest. Liquidity and structure: cash, undrawn committed facilities, maturities, covenants, ranking. Rating agencies publish sector methodologies that map these factors to rating outcomes; analysts reproduce the logic and form a view on whether the issuer will be upgraded or downgraded. For a steelmaker rated at the bottom of investment grade (BBB−), the critical question is the risk of becoming a fallen angel: a downgrade to high yield forces some investors to sell and typically widens spreads by much more than the change in default probability alone would justify (see C08).
Disclosure. Under the Market Abuse Regulation (MAR, Regulation (EU) No 596/2014, Article 20) and its Delegated Regulation (EU) 2016/958, anyone producing an investment recommendation must present it objectively, identify the author, distinguish facts from opinions and disclose conflicts of interest. A research report therefore states the analyst’s name, the date and time of production and first dissemination, the valuation method and its risks, the rating history for the last twelve months, the distribution of the broker’s ratings, and the relationships between the bank and the issuer: lead manager of an offering in the last twelve months, market maker, holdings above thresholds, investment banking fees.
Independence. The MiFID II Delegated Regulation (EU) 2017/565 (Articles 36 and 37) treats investment research as a source of conflicts and requires physical and information barriers, a ban on analysts taking part in pitches for investment banking mandates, a ban on promising favourable coverage, and limits on issuer review of drafts: the company may check facts before publication, but may not see the rating or target price. Analysts may not trade the securities they cover against their recommendations. In the United States, FINRA Rule 2241 and the 2003 Global Research Analyst Settlement impose comparable separation and forbid tying analysts’ pay to investment banking revenue; European banks distributing to US investors apply both regimes.
Payment: unbundling, then partial re-bundling. From 3 January 2018, MiFID II required EU investment firms to pay for research separately from execution, either from their own profit and loss account or through a research payment account funded by clients with a budget agreed in advance. Most asset managers chose to pay themselves, and broker research budgets and small-cap coverage both fell. A 2021 amendment allowed joint payment for research on issuers below EUR 1bn of market capitalisation; it was little used. The Listing Act (Directive (EU) 2024/2811) removed the threshold: firms may now pay for research jointly with execution, whatever the issuer, provided they agree a methodology for remunerating research with the provider, inform clients of their choice and assess the quality of the research every year. The deadline for national transposition was 5 June 2026, with application from 6 June 2026, and the technical rules were published in Delegated Directive (EU) 2026/374 (Official Journal of 2 June 2026). The Listing Act also created a framework for issuer-sponsored research, paid for by the company but produced under an ESMA code of conduct. The United Kingdom re-allowed bundled payment from August 2024; in the United States, the SEC’s temporary relief that let US brokers accept hard-currency research payments from EU clients expired in July 2023.
WHY IT MATTERS ON THE DESK — The payment model determines who pays for an ESG section in a steel report. If clients do not value it in their broker review, the analyst will not get the resources to build a carbon model. Re-bundling makes research budgets less visible and arguably favours large integrated brokers; whether it brings back coverage of mid-caps, where ESG data are thinnest, is still an open question in late 2026.
An ESG factor matters to a sell-side analyst only if it changes a number in the model. There are three places it can do so.
Cash flows. A carbon price, a fine, a lost licence, a price premium for low-carbon products, a change in demand. This is the most robust channel and the one the case uses: the analyst writes the cost or revenue into the forecast, with explicit assumptions.
The discount rate. A higher cost of equity or debt for exposed companies, either because their risk is higher (more volatile cash flows, policy tail risk) or because a growing share of investors refuse to hold them, which raises the return demanded by the remaining holders. This channel is real in theory and contested in size; adding an arbitrary “ESG premium” to the WACC is the classic way to double-count a risk already in the cash flows.
Terminal value and asset life. In capital-intensive sectors the terminal value is often 60% or more of enterprise value. Shortening the life of a blast furnace, adding a decommissioning provision or assuming that 2034 carbon costs persist forever changes the valuation more than any near-term forecast.
PITFALL — Do not adjust both the cash flows and the discount rate for the same risk. If the model already deducts the full expected carbon cost, adding 100 bp to the WACC “for climate risk” penalises the company twice. A discount rate adjustment is defensible only for what the cash flow scenarios do not capture, and it should be stated and sized.
Financial materiality asks whether an issue can affect the company’s cash flows, risk or cost of capital. Impact materiality asks whether the company affects people and the environment. Investment research is about the first; clients’ regulatory reporting (SFDR principal adverse impacts, for example) often needs data about the second.
The most used map of financial materiality is the set of 77 industry SASB Standards, maintained since 2022 by the International Sustainability Standards Board (ISSB) and referenced by IFRS S1 as a source of guidance on industry-specific disclosures. For iron and steel producers the standard lists greenhouse gas emissions, air emissions, energy management, water management, waste management, workforce health and safety and supply chain management (sourcing of iron ore and coking coal). Data vendors publish their own maps, which differ in weights. In the EU the European Sustainability Reporting Standards (ESRS) under the CSRD use double materiality, so companies in scope report both dimensions. The Omnibus I Directive, published in the Official Journal on 26 February 2026 and in force since 18 March 2026, narrowed CSRD scope to EU groups with more than 1,000 employees and more than EUR 450m of turnover; a listed steelmaker of the size in our case remains in scope, which gives the analyst audited emissions, a transition plan disclosure under ESRS E1 and taxonomy KPIs.
For a steelmaker, the material issue dominates all others: CO2. The sector emits roughly 1.9 tonnes of CO2 per tonne of crude steel on worldsteel’s 2024 figures (1.92 globally, 2.34 for the blast furnace-basic oxygen furnace route, 1.47 for gas-based direct reduction with electric arc furnace and 0.69 for scrap-based electric arc furnace, on worldsteel’s methodology, which includes electricity and some upstream emissions). It is the one ESG issue large enough to move a steel target price by tens of per cent.
EU ETS and free allocation. An EU steel plant surrenders one allowance (EUA) per tonne of CO2 it emits. It receives a free allocation based on product benchmarks and historical activity levels, fixed for each allocation period (2026-2030 is the current one) and adjusted only if activity changes significantly. The plant pays for the shortfall, emissions minus free allocation, at the market price. The December 2026 EUA contract traded around EUR 90 per tonne in early November 2026, according to ABN AMRO’s market commentary of 5 November 2026, with a forecast 2026 average of about EUR 80.
CBAM and the phase-out. The Carbon Border Adjustment Mechanism (Regulation (EU) 2023/956) entered its definitive period on 1 January 2026: importers of steel, aluminium, cement, fertilisers, hydrogen and electricity owe certificates for the embedded emissions of their imports, with certificate sales from 1 February 2027 and the first annual declaration due on 30 September 2027 for 2026 imports. In parallel, free allocation for CBAM sectors is reduced by the CBAM factor set in Article 10a(1a) of the ETS Directive: 97.5% of the benchmark allocation in 2026, 95% in 2027, 90% in 2028, 77.5% in 2029, 51.5% in 2030, 39% in 2031, 26.5% in 2032, 14% in 2033 and zero from 2034. The Commission’s ETS review proposal of 17 July 2026 (COM(2026) 616) would slow this schedule (59% in 2030, 15% from 2034 to 2037); it is under negotiation, and the law in force is the schedule above. A proposal for a Temporary Decarbonisation Fund (COM(2025) 990, December 2025) to support exporters exposed to leakage was also in the legislative process in 2026. For an analyst this means two moving parts with opposite signs: the company’s net carbon bill rises steeply after 2029, while CBAM raises the carbon cost of competing imports and, if it works, the EU steel price.
Trade defence. The EU steel safeguard expired on 30 June 2026 and was replaced from 1 July 2026 by a permanent measure, Regulation (EU) 2026/1384, which cut duty-free quotas by an average of 47% and raised the out-of-quota duty from 25% to 50%, with a “melt and pour” origin rule from 1 October 2026. For European steel margins this is at least as important as carbon in the near term, and the analyst has to separate the two effects when explaining a forecast change.
Pass-through. The share of a cost increase that a company recovers through higher prices. In steel it is a market outcome, set by the cost of the marginal supplier, not a company decision. If imports pay the full carbon cost at the border, domestic prices can rise by the carbon cost of the marginal import, which may be more or less than the company’s own carbon cost per tonne. Exports to markets without carbon pricing recover nothing.
Capex alignment. Under the EU Taxonomy, a non-financial company reports the share of its capital expenditure that is taxonomy-aligned, including capex under a plan to make an activity aligned. An analyst reads it as a signal of the transition, with two caveats: alignment for steel depends on technical screening criteria that a gas-based direct-reduction plant may not meet, and a high capex share says nothing about whether the capex earns its cost of capital.
Transition plan assessment. Analysts and their clients use frameworks such as the Transition Pathway Initiative (TPI) sector benchmarks, the Science Based Targets initiative’s steel guidance, the Climate Action 100+ net zero company benchmark and the disclosure requirements of ESRS E1 or IFRS S2. The case builds a scorecard in this tradition.
Rating agencies state that they integrate ESG factors where they are material to credit. Fitch publishes ESG relevance scores per issuer; Moody’s publishes issuer profile scores and an ESG credit impact score; S&P Global Ratings stopped publishing separate ESG credit indicators in 2023 and kept narrative ESG paragraphs in its reports. Credit research reads carbon through the same three channels as equity, with a shorter horizon and an asymmetric pay-off: bondholders gain little if a transition succeeds beyond the bond’s maturity, and lose a great deal if the capex cycle pushes leverage over a rating threshold before then. For green, sustainability-linked and transition bonds (see C01 and C02) the credit analyst also comments on the framework and the use of proceeds, but the credit view is set by the issuer, not the label.
Analysts buy ESG data from a small number of large providers (MSCI, Morningstar Sustainalytics, ISS ESG, S&P Global, LSEG, Bloomberg) and from specialist sources (CDP questionnaires, SBTi’s target dashboard, TPI). Ratings from different providers agree much less than credit ratings do: Berg, Kölbel and Rigobon (2022) found correlations between six providers’ scores averaging about 0.54, against near-perfect agreement among credit ratings, and traced most of the gap to differences in what is measured and how. Estimated emissions also differ by vendor when companies do not disclose (see A03).
The ESG Ratings Regulation (Regulation (EU) 2024/3005) applies from 2 July 2026. Providers of ESG ratings in the EU must be authorised and supervised by ESMA, publish their methodologies and separate rating activity from certain other services; providers already active had to notify ESMA by 2 August 2026 and apply by 2 November 2026 to continue under the transitional regime. For a bank research department two questions follow: whether any score it publishes is an “ESG rating” within the Regulation’s definition, which the bank’s legal team has to settle given the exclusions for regulated activities, and what it must disclose when it quotes a third-party ESG rating in a report. The cautious practice in 2026 is to present ESG analysis as an input to the investment view rather than as a stand-alone score sold to clients.
Clients ask the question constantly, and research teams publish back-tests. The academic evidence teaches caution.
Expected and realised returns are different things. Pástor, Stambaugh and Taylor (2021, 2022) show that if many investors prefer green assets, green stocks should have lower expected returns in equilibrium, yet they can deliver higher realised returns for years while climate concerns rise unexpectedly and money flows into them. Their decomposition attributes the strong performance of green US stocks in the 2010s largely to such unexpected shifts. A back-test of a period of rising concern therefore says little about the future.
The carbon premium. Bolton and Kacperczyk (2021) report higher returns for stocks with higher emissions, consistent with investors demanding compensation for carbon risk. Aswani, Raghunandan and Rajgopal (2024) find that the result depends on vendor-estimated emissions and on unscaled measures, and largely disappears with disclosed emissions and intensity. The debate is about data as much as about finance.
Credit. Seltzer, Starks and Zhu find that firms with poor environmental profiles have lower ratings and higher yields, especially where climate regulation is stricter; Ilhan, Sautner and Vilkov (2021) find that carbon-intensive firms carry more downside tail risk in option prices. These are risk results, not alpha.
Meta-analyses. Friede, Busch and Bassen (2015) count about 90% of studies with a non-negative relation between ESG and financial performance, but the studies vary widely in quality and definition. Roncalli’s handbook, chapter 3, summarises the professional consensus as dynamic: the relation between ESG and performance depends on period, region, implementation and flows, and a study from one decade does not carry over to the next.
Statistical power. The most underrated point. A long-short ESG strategy with a true premium of 2% a year and a tracking volatility of 8% a year needs decades of data before its t-statistic reliably exceeds 2.
FORMULA
is the t-statistic of the average annual excess return; the true annual premium; the annual volatility of the long-short return; the number of years of data; the years needed to reach a target t-statistic (2 by convention, 3 under the stricter hurdle proposed by Harvey, Liu and Zhu, 2016, for factors found by data mining).
INTERACTIVE LAB — How many years of data does it take to prove an ESG premium?
Show that a plausible ESG or carbon premium cannot be distinguished from noise over the few years of reliable ESG data available.
| Parameter you can move | Range | Default |
|---|---|---|
| True annual premium | 0.5 to 6 % | 2 % |
| Volatility of the long-short return | 2 to 20 % | 8 % |
| Years of data available | 3 to 60 years | 15 years |
| Result at default settings | Value |
|---|---|
| Expected t-statistic | 0.97 |
| Years needed for t = 2 | 64 years |
| Years needed for t = 3 | 144 years |
Try it. Keep a 2% premium and lower volatility to 4%: 16 years are still needed. Then ask how many years of comparable ESG scores exist.
The desk lesson: an analyst should present ESG back-tests with their t-statistics and periods, distinguish risk from return, and prefer the cash-flow channel, where the economics can be checked, over claims about alpha.
Research has no deal in the banking sense; its equivalent is the initiation of coverage, which takes eight to fourteen weeks from the decision to cover a company to the first report, and then the maintenance of coverage for years.
INTERACTIVE LAB — Initiating coverage of a steelmaker, from coverage request to the first client meeting
Follow an initiation of coverage through the research department, its controls and its distribution.
| # | Who | Does what | Deliverable |
|---|---|---|---|
| 1 | Head of research and research management committee | Approve the decision to cover Acierna Group, based on client demand, sales feedback and team capacity; check that no banker request drives the choice | Coverage decision minuted |
| 2 | Control room | Check the watch and restricted lists; confirm that Meridian holds no inside information through the bank and that no offering by Acierna is in progress | Clearance to initiate |
| 3 | Lead metals and mining analyst and associate | Build the earnings model from filings, sector data and management meetings held in public-side conditions | Draft model and investment thesis |
| 4 | ESG and climate research specialist | Supply the carbon dataset: verified ETS emissions, free allocation, CBAM schedule, carbon price deck, transition plan disclosures; challenge the pass-through assumption | Carbon module of the model and ESG section draft |
| 5 | Credit research analyst | Build the credit view on the same model: adjusted leverage, FFO to debt, liquidity, bond curve, rating threshold analysis | Credit initiation draft and bond recommendations |
| 6 | Lead analyst | Set the valuation, scenarios, rating and target price; write the summary and the risks to the rating | Full draft of the initiation |
| 7 | Research review committee (senior analysts and head of research) | Challenge the thesis, the scenario weights and the consistency of equity and credit views | Approved investment view |
| 8 | Acierna investor relations | Check factual sections only, without the rating, target price or valuation, as allowed by the independence rules | Factual corrections |
| 9 | Research compliance (supervisory analyst) | Review disclosures, conflicts, rating distribution, ESG claims and data sources; check the report against MAR and MiFID requirements | Publication approval |
| 10 | Publishing and distribution | Release the report simultaneously to all entitled clients through the research portal and aggregators; record time of dissemination | Published initiation |
| 11 | Equity and credit sales | Present the call at the morning meeting and to clients; collect feedback | Client call list and feedback log |
| 12 | Lead analyst with corporate access | Market the report on a roadshow to portfolio managers and stewardship teams; organise a field trip to the plant | Meetings logged for the broker review |
Coverage decision. Coverage is a resource decision with conflict risk: an analyst team covers 15 to 25 companies, and adding one means dropping or neglecting another. Requests from bankers are allowed as information about client interest but cannot decide coverage. The decision is minuted precisely so that the bank can show, if challenged, that coverage was not promised to an issuer.
Clearance and the information barrier. The control room checks whether the bank is working on a confidential transaction for Acierna or its competitors. If Meridian were mandated on a capital increase, the initiation would be postponed and subject to the quiet periods of the offering (see C05). Analysts can be wall-crossed only through the control room, with a chaperone, and cannot publish on the company while they hold inside information.
Model building. The associate starts from annual reports, the sustainability statement, ETS registry data (installation-level verified emissions and free allocation are public through the EU Union Registry) and industry statistics. Management meetings follow the selective disclosure rules: the company may explain public information but must not give the analyst non-public information. The ESG specialist owns the carbon price deck, which must be the house deck shared by all sector analysts, so that a cement and a steel company are not valued at different carbon prices in the same week.
Credit and equity on the same model. Meridian runs one operating model per company and lets the equity and credit analysts draw different conclusions. The review committee checks that differences are explained by horizon and pay-off, not by inconsistent forecasts.
Issuer review and compliance. The issuer may correct facts in the draft but never sees the conclusion. Research compliance then checks disclosures, conflicts and claims. In 2026 compliance teams pay particular attention to ESG wording: a report that calls a company “green” or “Paris-aligned” without a stated method exposes the bank to greenwashing complaints under the same principles that ESMA and the other European Supervisory Authorities set out for market participants in 2024.
Distribution and marketing. All clients entitled to the research receive it at the same time. The analyst then markets the report in meetings, which is where most of its value is created and where clients judge its usefulness in their annual broker review.
Question. Meridian’s metals and mining team has been asked by several clients to initiate on Acierna Group, a listed European integrated steelmaker. The investment question is simple to state: the stock trades at a low multiple, consensus EBITDA is flat, and the company has announced a transition plan. Is the stock cheap, or is the market rightly discounting carbon?
The team: Léa, the lead metals and mining analyst, owns the call; Tomás, her associate, builds the model; Sofia, from the ESG and climate research team, owns the carbon module; Malik, in credit research, covers Acierna’s bonds. Consensus, compiled by a data vendor from about fifteen brokers, shows 2027 EBITDA of roughly EUR 1.05bn, flat on 2026, and a majority of Buy ratings. Only a minority of the brokers publish an explicit carbon line. The initiation will be read by three kinds of client: generalist European equity funds asking whether the stock is a value opportunity; credit funds holding Acierna’s 2029 and 2032 bonds; and stewardship teams preparing their engagement on the transition plan before the 2027 annual general meeting.
Data (case assumptions). All figures are for the 2026 financial year unless stated, in EUR.
| Item | Value |
|---|---|
| Crude steel output | 10.0 Mt: 7.0 Mt blast furnace (three furnaces), 3.0 Mt scrap EAF |
| Shipments | 9.4 Mt, of which 75% in the EU, 25% exported |
| Direct (scope 1) intensity | BF-BOF 1.90 t CO2/t; scrap EAF 0.08; DRI-EAF on gas 0.60 |
| Scope 2 intensity | BF-BOF 0.15; scrap EAF 0.30; DRI-EAF 0.40 |
| Free allocation before CBAM factor | 11.4 Mt a year, held flat to 2034 |
| EUA price deck | EUR 80 in 2026, +EUR 7.5 a year (EUR 110 in 2030, 140 in 2034) |
| EBITDA before carbon and CBAM effects | EUR 1,000m, mid-cycle, held flat in real terms |
| Net debt; pensions | EUR 1,800m; EUR 300m |
| Shares; share price | 200m; EUR 11.80 |
| Credit rating | BBB−, stable (two agencies) |
The price deck sits below the December 2026 futures price of about EUR 90 at the time of writing, because the house deck uses the average expected for 2026 and a linear rise after; the ESG specialist tests higher paths in the scenarios. Holding pre-carbon EBITDA flat isolates the carbon effect: everything that follows measures what carbon and CBAM do to an otherwise unchanged business.
CASE — The analyst has two separate jobs: measure what carbon costs Acierna on current policy, which is a forecasting question; and judge whether the transition plan will be delivered, which is an assessment of probabilities. Stages 5.2 to 5.4 do the first, 5.5 to 5.7 the second.
Question. How much does Acierna pay for carbon today, and in 2030 and 2034 if nothing changes?
Calculation. Emissions are the sum over routes of output times direct intensity. Free allocation is the base allocation times the CBAM factor of the year. The net carbon cost is the shortfall times the price.
FORMULA
is the net carbon cost in year (EUR m); verified scope 1 emissions (Mt CO2); the free allocation before the CBAM factor (11.4 Mt); the CBAM factor (97.5% in 2026 to zero in 2034); the EUA price (EUR per tonne).
Emissions are 7.0 × 1.90 + 3.0 × 0.08 = 13.54 Mt. In 2026 free allocation is 11.4 × 97.5% = 11.115 Mt, the shortfall 2.425 Mt and the net cost 2.425 × 80 = EUR 194m. The free allowances themselves are worth EUR 889.2m at the 2026 price. In 2030 the CBAM factor falls to 51.5%: free allocation drops to 5.871 Mt, the shortfall rises to 7.669 Mt and, at EUR 110, the bill reaches EUR 843.6m. In 2034, with no free allocation at EUR 140, it is EUR 1,895.6m, 9.8 times the 2026 bill.
Sourcing the data. Tomás does not take emissions from a vendor. Verified emissions and free allocation per installation are public in the EU Union Registry and reconcile with the ESRS E1 disclosure in Acierna’s sustainability statement; vendor figures are used only to cross-check. The base allocation of 11.4 Mt is held flat because allocations for 2026-2030 are fixed on historical activity levels and change only if activity at an installation moves by more than 15%; after 2030 the analyst assumes the same level, which is a simplification the report states. Benchmark updates for the 2031-2035 period would lower it, so the assumption is if anything generous to the company.
INTERACTIVE LAB — Acierna’s net carbon cost and the value of its free allocation, 2026-2034
Compute the net carbon bill year by year under the CBAM factor schedule in force, and see how much of today’s economics rests on free allocation.
| Parameter you can move | Range | Default |
|---|---|---|
| EUA price in 2026 | 50 to 150 EUR/t | 80 EUR/t |
| Annual increase in the EUA price | 0 to 20 EUR/t | 7.5 EUR/t |
| Free allocation before CBAM factor | 8 to 14 Mt | 11.4 Mt |
| Blast furnace output | 0 to 10 Mt | 7 Mt |
| Blast furnace direct intensity | 1.5 to 2.3 t CO2/t | 1.9 t CO2/t |
| Scrap EAF output | 0 to 10 Mt | 3 Mt |
| Scrap EAF direct intensity | 0 to 0.3 t CO2/t | 0.08 t CO2/t |
| Result at default settings | Value |
|---|---|
| Scope 1 emissions | 13.54 Mt CO2 |
| Free allocation 2026 | 11.115 Mt |
| Shortfall 2026 | 2.425 Mt |
| Net carbon cost 2026 | 194.0 EUR m |
| Value of free allocation 2026 | 889.2 EUR m |
| Free allocation 2030 | 5.871 Mt |
| Shortfall 2030 | 7.669 Mt |
| Net carbon cost 2030 | 843.6 EUR m |
| Net carbon cost 2034 | 1,895.6 EUR m |
| 2034 bill as a multiple of 2026 | 9.8 x |
Try it. Set the price flat at EUR 90: the 2034 bill is still EUR 1.2bn, because the phase-out of free allocation, not the price, drives most of the increase.
Reading. The free allocation is worth EUR 889.2m in 2026, close to the company’s whole EBITDA (see 5.3). Today’s profits rest on a subsidy that the law in force removes in eight years. That is the first sentence of the ESG section of the initiation, and the reason why carbon is not a footnote in this report.
PITFALL — Do not compute the carbon cost as emissions times price and subtract it from reported EBITDA. Reported EBITDA already bears the net cost of the shortfall (EUR 194m in 2026); subtracting the gross cost again double-counts it. The right baseline is EBITDA before carbon, from which each year’s net cost is deducted.
Question. How much of the rising bill can Acierna recover through prices, and what EBITDA does that leave?
Reasoning. Under CBAM an importer pays for the embedded emissions of its steel minus the free allocation an EU producer of the same product would receive. If the marginal import comes from a blast furnace at 1.90 t CO2 per tonne, its carbon cost per tonne is (1.90 − 1.40 × CF) × P, where 1.40 t is the benchmark-equivalent free allocation per tonne (case assumption). If EU prices rise by a share θ of that cost, Acierna gains the uplift on its EU shipments and nothing on its exports.
FORMULA
is the EU price uplift per tonne (EUR); the share of the marginal import’s carbon cost reflected in EU prices (80%, case assumption); the embedded intensity of the marginal import (1.90 t CO2/t); the benchmark-equivalent free allocation per tonne (1.40); EU shipments (9.4 × 75% = 7.05 Mt); EBITDA before carbon and CBAM effects (EUR 1,000m); the net carbon cost of 5.2.
In 2026 the uplift is 0.8 × (1.90 − 1.365) × 80 = EUR 34.24 per tonne, worth EUR 241.4m on 7.05 Mt. Reported EBITDA is 1,000 + 241.4 − 194.0 = EUR 1,047.4m. The same business paying for all its emissions without free allocation would earn 1,000 + 241.4 − 13.54 × 80 = EUR 158.2m: the full carbon cost EBITDA, the figure some ESG analysts publish to show dependence on free allocation.
In 2030 the uplift rises to EUR 103.75 per tonne (EUR 731.5m), against a net carbon cost of EUR 843.6m: Acierna recovers 86.7% of its bill. Carbon-adjusted EBITDA is EUR 887.9m, 15.2% below 2026. By 2034 it is EUR 604.6m, 42.3% below.
INTERACTIVE LAB — Carbon-adjusted EBITDA with market-driven pass-through
Combine the net carbon bill with the EU price uplift created by CBAM and compare carbon-adjusted EBITDA with a carbon-blind forecast.
| Parameter you can move | Range | Default |
|---|---|---|
| Share of import carbon cost reflected in EU prices | 0 to 100 % | 80 % |
| Embedded intensity of the marginal import | 1 to 2.5 t CO2/t | 1.9 t CO2/t |
| Benchmark-equivalent free allocation per tonne | 1 to 1.6 t CO2/t | 1.4 t CO2/t |
| Shipments | 5 to 15 Mt | 9.4 Mt |
| EU share of shipments | 0 to 100 % | 75 % |
| EBITDA before carbon and CBAM effects | 500 to 2000 EUR m | 1000 EUR m |
| EUA price in 2026 | 50 to 150 EUR/t | 80 EUR/t |
| Annual increase in the EUA price | 0 to 20 EUR/t | 7.5 EUR/t |
| Free allocation before CBAM factor | 8 to 14 Mt | 11.4 Mt |
| Scope 1 emissions (from 5.2) | 5 to 20 Mt | 13.54 Mt |
| Result at default settings | Value |
|---|---|
| EU shipments | 7.05 Mt |
| EU price uplift 2026 | 34.24 EUR/t |
| Uplift revenue 2026 | 241.4 EUR m |
| Reported EBITDA 2026 | 1,047.4 EUR m |
| Full carbon cost EBITDA 2026 | 158.2 EUR m |
| EU price uplift 2030 | 103.75 EUR/t |
| Uplift revenue 2030 | 731.5 EUR m |
| Net carbon cost 2030 | 843.6 EUR m |
| Share of 2030 carbon bill recovered | 86.7 % |
| Own net carbon cost per tonne shipped 2030 | 89.7 EUR/t |
| Carbon-adjusted EBITDA 2030 | 887.9 EUR m |
| Change vs 2026 | -15.2 % |
| Carbon-adjusted EBITDA 2034 | 604.6 EUR m |
| Change 2034 vs 2026 | -42.3 % |
Try it. Set the pass-through share to 100%: 2034 EBITDA recovers to about EUR 980m, close to today. The whole equity story hinges on one parameter that nobody can observe in advance.
Reading. Two lessons. First, a cleaner-than-marginal producer can recover more than its own carbon cost per tonne on its domestic sales: in 2030 the uplift (EUR 103.75 per tonne) exceeds Acierna’s own net cost per tonne shipped (EUR 89.7). Second, exports earn no uplift and the company’s own emissions are high, so with θ at 80% the bill still outruns the uplift. The analyst must show θ explicitly and the sensitivity to it; consensus models that keep EBITDA flat are implicitly assuming θ close to 100% and ignoring the export leak.
The sensitivity is large. With θ at 100%, 2034 EBITDA is about EUR 980m, almost unchanged from today; at 80% it is EUR 604.6m. Sofia sets 80% as the house case for three reasons she writes into the report: default values and verification of import emissions will take time to bite; some imports come from lower-carbon routes and set lower prices; and the new trade measure reduces import volumes, which makes the marginal tonne a domestic one, whose price is set by EU competitors with their own free allocation. She also records what would change her view: evidence from 2027 price data that the EU-world steel price spread has widened by roughly the CBAM cost.
PITFALL — Pass-through is not a company characteristic. Writing “Acierna passes through 60% of its carbon costs” invites the reader to think the company sets it. The uplift is set by the cost of the marginal tonne in the EU market, which depends on CBAM enforcement, default values for imports, the trade measure of July 2026 and demand. A stronger CBAM helps Acierna even if it does nothing itself.
Question. Acierna’s plan converts two blast furnaces to gas-based direct reduction with electric arc furnaces: 2.5 Mt from 1 January 2030 and 2.5 Mt from 1 January 2033, leaving one 2.0 Mt blast furnace. What does it change?
Data (case assumptions). Each phase costs EUR 2,100m, spent over three years (500, 800, 800: 2027-2029 for phase 1, 2030-2032 for phase 2), with public grants of 25%. The converted capacity runs at 0.60 t CO2 per tonne directly and costs EUR 40 per tonne more to operate than the blast furnace (gas and pellet premium net of coke savings). Maintenance capex is EUR 350m a year in both scenarios.
Calculation. Emissions fall to 4.5 × 1.90 + 2.5 × 0.60 + 0.24 = 10.29 Mt in 2030 and to 7.04 Mt from 2033, 48.0% below 2026. Carbon-adjusted EBITDA with the plan is EUR 1,145.4m in 2030, against EUR 887.9m without, and EUR 1,314.6m in 2034, against EUR 604.6m: a gap of EUR 710m a year at the end of the forecast.
Transition capex of EUR 4,200m represents 66.7% of total capex over 2027-2032 (4,200 over 4,200 plus six years of maintenance at 350). The abatement cost of phase 1 is the annualised net capex plus the extra opex, divided by the tonnes abated:
FORMULA
is the abatement cost (EUR per tonne of CO2); the capex of the phase (EUR 2,100m); the grant share (25%); the annuity factor at the WACC (9%) over the asset life (20 years); the extra operating cost per tonne of steel (EUR 40); converted capacity (2.5 Mt); and the direct intensities (1.90 and 0.60).
The annual capital charge is EUR 172.5m; with EUR 100m of extra opex, the 3.25 Mt abated cost EUR 83.9 per tonne, below the EUA price from 2027 in the house deck.
INTERACTIVE LAB — Transition plan: emissions, carbon-adjusted EBITDA with and without the plan, abatement cost
Measure what the two-phase DRI-EAF plan changes in emissions, EBITDA and capex mix, and what each tonne abated costs.
| Parameter you can move | Range | Default |
|---|---|---|
| Capacity converted per phase | 1 to 3.5 Mt | 2.5 Mt |
| DRI-EAF direct intensity | 0.1 to 1.2 t CO2/t | 0.6 t CO2/t |
| Extra operating cost of DRI-EAF | 0 to 120 EUR/t | 40 EUR/t |
| Capex per phase | 1000 to 4000 EUR m | 2100 EUR m |
| Grant share | 0 to 50 % | 25 % |
| Phase 2 executed (2033) | yes / no | yes |
| Share of import carbon cost reflected in EU prices | 0 to 100 % | 80 % |
| EUA price in 2026 | 50 to 150 EUR/t | 80 EUR/t |
| Annual increase in the EUA price | 0 to 20 EUR/t | 7.5 EUR/t |
| Result at default settings | Value |
|---|---|
| Emissions 2030 with plan | 10.29 Mt |
| Emissions 2034 with plan | 7.04 Mt |
| Reduction 2034 vs 2026 | 48.0 % |
| Carbon-adjusted EBITDA 2030, plan | 1,145.4 EUR m |
| Carbon-adjusted EBITDA 2030, no plan | 887.9 EUR m |
| Carbon-adjusted EBITDA 2034, plan | 1,314.6 EUR m |
| EBITDA gap 2034, plan minus no plan | 710.0 EUR m |
| Transition share of 2027-2032 capex | 66.7 % |
| Annual capital charge of phase 1 | 172.5 EUR m |
| Tonnes abated by phase 1 | 3.25 Mt |
| Abatement cost of phase 1 | 83.9 EUR/t CO2 |
Try it. Raise the extra operating cost to EUR 100 per tonne (a gas price shock): the abatement cost rises to about EUR 130 per tonne, close to the 2033 carbon price, and the plan’s advantage shrinks sharply.
Reading. The plan is economically rational on the house deck: it abates at EUR 83.9 per tonne when carbon costs EUR 110 to 140. But it is a gas-based plan. Its abatement cost depends on the gas price, its emissions of 0.60 t per tonne remain far from near-zero steel, and whether its capex counts as taxonomy-aligned is uncertain. The analyst reports the 66.7% as transition capex, not as aligned capex, and says why.
Three further questions go into the report because clients will ask them. Is the plan locked into gas? A gas-based direct-reduction shaft can in principle run on hydrogen later, which is the argument the company makes; the analyst notes that hydrogen at the volumes needed is neither contracted nor priced, so the model gives no value to the switch, and shows separately what it would take: at 0.10 t per tonne the phase-1 abatement cost would fall, but only if hydrogen costs less than the gas and carbon it replaces. Is the opex penalty right? EUR 40 per tonne reflects a gas price assumption; Sofia tests a range, and the try-it of the block shows that the plan’s economics are more sensitive to gas than to carbon. What do peers do? Several European producers have cut or delayed hydrogen-ready plans since 2024 and favoured electric arc furnaces fed with scrap and purchased metallics; the report presents Acierna’s plan as more ambitious than some peers and therefore more exposed to execution risk, rather than as best in class.
PITFALL — A high transition capex share is not proof of value creation. The question for the equity analyst is whether the plan earns its cost of capital, which depends on carbon prices, pass-through and gas prices; the question for the credit analyst is whether the balance sheet can carry the spending. Both are answered in the next stages, not by the percentage.
Question. What is Acierna worth with and without the plan, and what does the market price imply?
Method. A DCF of free cash flow to the firm, 2027 to 2034, with a terminal value on 2034 cash flow at zero growth, discounted at a WACC of 9%. Free cash flow is EBITDA minus 25% tax on EBITDA less depreciation of EUR 350m, minus maintenance capex of EUR 350m and transition capex net of grants. Three cases:
INTERACTIVE LAB — Target price by scenario, probability-weighted target price and market-implied probability of delivery
Value Acierna under three cases, weight the plan by the analyst’s probability of delivery, and back out the probability the share price implies.
| Parameter you can move | Range | Default |
|---|---|---|
| WACC | 6 to 12 % | 9 % |
| Probability that the plan is delivered | 0 to 100 % | 60 % |
| Current share price | 3 to 30 EUR | 11.8 EUR |
| Share of import carbon cost reflected in EU prices | 0 to 100 % | 80 % |
| EUA price in 2026 | 50 to 150 EUR/t | 80 EUR/t |
| Annual increase in the EUA price | 0 to 20 EUR/t | 7.5 EUR/t |
| Capacity converted per phase | 1 to 3.5 Mt | 2.5 Mt |
| DRI-EAF direct intensity | 0.1 to 1.2 t CO2/t | 0.6 t CO2/t |
| Extra operating cost of DRI-EAF | 0 to 120 EUR/t | 40 EUR/t |
| Capex per phase | 1000 to 4000 EUR m | 2100 EUR m |
| Grant share | 0 to 50 % | 25 % |
| Phase 2 executed (2033) | yes / no | yes |
| Net debt | 0 to 4000 EUR m | 1800 EUR m |
| Pension deficit | 0 to 1000 EUR m | 300 EUR m |
| Shares outstanding | 100 to 400 m | 200 m |
| Result at default settings | Value |
|---|---|
| EV, carbon-blind | 5,812 EUR m |
| Implied EV/EBITDA 2026, carbon-blind | 5.55 x |
| Target price, carbon-blind | 18.56 EUR |
| Upside, carbon-blind | 57.3 % |
| EV, no plan | 3,282 EUR m |
| Target price, no plan | 5.91 EUR |
| EV, plan | 4,876 EUR m |
| Target price, plan | 13.88 EUR |
| Terminal value share of EV, plan | 82.7 % |
| Probability-weighted target price | 10.69 EUR |
| Upside to weighted target price | -9.4 % |
| Rating (1 Buy, 0 Hold, -1 Sell) | 0 |
| Probability of delivery implied by the share price | 73.9 % |
Try it. Set the pass-through share to 100%: the no-plan value jumps and the gap between scenarios narrows. The plan is partly a hedge against CBAM failing to lift EU prices.
Results. The carbon-blind DCF gives an enterprise value of EUR 5,811.6m (5.55 times 2026 EBITDA) and a target price of EUR 18.56, 57.3% above the share price: a screaming Buy that explains why a carbon-blind consensus can look cheap for years. With carbon, the no-plan case is worth EUR 5.91 a share and the plan case EUR 13.88. The terminal value is 82.7% of the plan EV, which tells the reader where the uncertainty lies.
Choices the analyst must defend. The WACC of 9% is the same in all three scenarios: the carbon risk is in the cash flows, and adding a premium to the discount rate would count it twice. The terminal value assumes that 2034 conditions persist, with a carbon price of EUR 140 and no free allocation; that is harsh for the no-plan case if policy softens and generous for the plan case if hydrogen becomes necessary after 2035. The model has no explicit green premium for low-carbon steel: some buyers in the automotive sector pay one, but volumes and levels are not disclosed reliably, so Léa treats it as upside rather than base case. Finally, the DCF ignores the ETS review proposal of July 2026, which would slow the phase-out; she shows it as a scenario in an appendix, because a proposal under negotiation is not law.
Weighting. The analyst assigns a 60% probability to delivery of the plan, informed by the scorecard of 5.7. The weighted target price is 0.6 × 13.88 + 0.4 × 5.91 = EUR 10.69, 9.4% below the share price: a Hold. Read backwards, the share price of EUR 11.80 implies a 73.9% probability of delivery. The investment debate becomes concrete: the analyst is more sceptical than the market about delivery, which is a different claim from the stock being “cheap”.
WHY IT MATTERS ON THE DESK — Clients do not need another target price; they need to know what the price assumes. “The market prices a 74% chance that Acierna delivers its plan; we think 60%, because phase 2 depends on gas, grants and permits” is a statement a portfolio manager can disagree with, test and trade on. It is also defensible in front of research compliance, because each number is traceable.
PITFALL — Weighting two scenarios is not the same as valuing one scenario with an average of the inputs. The no-plan scenario has a different capex path and a different terminal cash flow; averaging the EBITDA paths and then valuing would understate the spread of outcomes, which is the information the reader needs.
Question. Can Acierna stay investment grade while it spends EUR 4.2bn?
Method. Net debt evolves with free cash flow after interest at 4.5% on opening net debt, cash taxes, capex and a dividend of EUR 150m. Meridian credit research reads the agencies’ published steel methodologies as requiring, for BBB−, adjusted debt (net debt plus pension deficit) of no more than 2.5 times EBITDA through the cycle and FFO of at least 30% of adjusted debt; these thresholds are the analyst’s reading, a case assumption.
INTERACTIVE LAB — Adjusted leverage with and without the transition plan against the BBB- threshold
Project adjusted net debt to EBITDA and FFO to debt through the capex cycle and identify the years of rating pressure.
| Parameter you can move | Range | Default |
|---|---|---|
| Annual dividend | 0 to 400 EUR m | 150 EUR m |
| Grant share | 0 to 50 % | 25 % |
| Capex per phase | 1000 to 4000 EUR m | 2100 EUR m |
| Phase 2 executed (2033) | yes / no | yes |
| Share of import carbon cost reflected in EU prices | 0 to 100 % | 80 % |
| EUA price in 2026 | 50 to 150 EUR/t | 80 EUR/t |
| Annual increase in the EUA price | 0 to 20 EUR/t | 7.5 EUR/t |
| Capacity converted per phase | 1 to 3.5 Mt | 2.5 Mt |
| DRI-EAF direct intensity | 0.1 to 1.2 t CO2/t | 0.6 t CO2/t |
| Extra operating cost of DRI-EAF | 0 to 120 EUR/t | 40 EUR/t |
| Opening net debt | 0 to 4000 EUR m | 1800 EUR m |
| Pension deficit | 0 to 1000 EUR m | 300 EUR m |
| Interest rate on net debt | 2 to 8 % | 4.5 % |
| Leverage threshold for BBB- | 1.5 to 4 x | 2.5 x |
| Modified duration of the 2032 bond | 2 to 9 years | 5.5 years |
| Spread widening on a downgrade to BB+ | 20 to 200 bp | 80 bp |
| Result at default settings | Value |
|---|---|
| Adjusted leverage 2026 | 2.00 x |
| Peak leverage with plan | 3.17 x |
| Year of peak | 2,032 |
| Years above threshold with plan | 3 years |
| Lowest FFO to adjusted debt with plan | 23.4 % |
| Leverage 2030 with plan | 2.47 x |
| Leverage 2030 without plan | 1.09 x |
| Leverage 2034 with plan | 1.83 x |
| Leverage 2034 without plan | 0.84 x |
| Approximate bond price change on downgrade | -4.4 points |
Try it. Set the dividend to zero: leverage never crosses 2.5x. The credit analyst’s recommendation to bondholders is, in effect, a recommendation on the dividend.
Results. Adjusted leverage starts at 2.00x. With the plan it crosses 2.5x in three years, peaks at 3.17x in 2032 with FFO to debt falling to 23.4%, then falls to 1.83x in 2034 as phase 2 starts and spending stops. Without the plan, leverage falls steadily to 0.84x in 2034, because no capex is spent while EBITDA erodes slowly.
Reading. The credit and equity analysts reach opposite intuitions from the same model. For a bondholder with a 2030 or 2032 maturity, the no-plan path looks safer: the damage from carbon arrives mostly after the bond is repaid. For a shareholder, the no-plan path destroys value. The credit initiation therefore recommends underweight on the 2031-2033 maturities, which carry the downgrade risk of the capex peak (a downgrade to BB+ with 80 bp of widening would cost about 4.4 points on a bond with a duration of 5.5), and market weight on short maturities, and states the remedy: a dividend cut or a hybrid issue in 2028-2029 would keep the issuer investment grade. A transition bond to finance the DRI plants (see C02) would not change this analysis; the label does not lower leverage.
Malik adds two items that equity readers often skip. Liquidity: Acierna has EUR 900m of cash and an undrawn EUR 1.5bn revolving credit facility maturing in 2030, so the capex peak is fundable even if the bond market is shut for a few months. Covenants: the bonds have none beyond a negative pledge and change-of-control clause, which is standard for investment grade, so bondholders have no protection against a dividend that the board chooses to maintain.
Question. How credible is the plan, and how should the 60% probability be justified?
Meridian’s scorecard has five pillars, each scored 0 to 4, with weights reflecting their link to delivery: targets and ambition (20%), capex alignment (25%), intensity against the reference pathway (25%), delivery dependencies such as gas, hydrogen, grid, permits and grants (15%), and governance and disclosure (15%). Two pillars are computed: capex alignment scores 4 × min(share / 80%, 1) on the 66.7% transition share of 5.4; intensity scores 2 plus one point for each 0.1 t of headroom below Meridian’s steel reference pathway of 1.75 t CO2 per tonne in 2030 (scopes 1 and 2), bounded between 0 and 4. The three others are judgements: targets 2 (a 2030 intensity target, no validated long-term target), delivery 1 (phase 2 not yet funded, gas exposure), governance 3 (board oversight, a carbon metric in executive pay, audited ESRS E1 disclosure).
INTERACTIVE LAB — Transition scorecard: company average versus primary-route comparison
Compute the transition score and see how the choice of intensity comparison changes the conclusion for a mixed-route steelmaker.
| Parameter you can move | Range | Default |
|---|---|---|
| Targets and ambition score | 0 to 4 0-4 | 2 0-4 |
| Delivery dependencies score | 0 to 4 0-4 | 1 0-4 |
| Governance and disclosure score | 0 to 4 0-4 | 3 0-4 |
| Intensity compared on | primary route / company | primary route |
| Reference pathway 2030 (scopes 1 and 2) | 1 to 2.2 t CO2/t | 1.75 t CO2/t |
| Transition share of capex (from 5.4) | 0 to 100 % | 66.67 % |
| Result at default settings | Value |
|---|---|
| Company intensity 2026 | 1.549 t CO2/t |
| Company intensity 2030 with plan | 1.286 t CO2/t |
| Primary-route intensity 2030 with plan | 1.675 t CO2/t |
| Intensity score | 2.75 0-4 |
| Capex alignment score | 3.33 0-4 |
| Transition score | 63.0 /100 |
| Grade (4 advanced, 3 credible with delivery risk, 2 partial, 1 lagging) | 3 |
Try it. Switch to the company-average comparison: the intensity pillar jumps to the maximum because 30% of output already comes from scrap EAFs, not because of the plan.
Results. Company intensity is 1.549 t per tonne in 2026 and 1.287 in 2030 with the plan, far below the 1.75 pathway, so the company-average comparison gives the maximum intensity score and a total of 70.8. The analyst compares instead the primary route (blast furnace and DRI only): 1.675 in 2030, just under the pathway, scoring 2.75, for a total of 63.0, grade “credible with delivery risk”.
Reading. A sector-average pathway rewards any producer with a high scrap share, whatever it does with its blast furnaces; scrap availability is limited, so a company cannot decarbonise the sector by owning more EAFs. Comparing the primary route is a choice the analyst must state. The score does not compute the 60% probability; it supports it: credible plan, strong governance, weak delivery pillar concentrated on phase 2. The 60% sits roughly where a 63/100 plan with an unfunded second phase belongs, and the report says that this is judgement.
Sofia cross-checks the result against external frameworks before publication. The Transition Pathway Initiative assesses management quality and carbon performance of listed steel companies on public disclosures, and the analyst checks that the scorecard does not contradict it without explanation. The scorecard is also shared, through sales, with stewardship teams preparing engagement: they want to know which pillar is weakest, because that is the question to ask the chair. Here the answer is clear: delivery, and specifically the funding of phase 2. A good scorecard ends in an engagement question, not only in a number.
PITFALL — Scorecards look objective because they produce a number. Their weights, thresholds and comparison choices are opinions. A report should show the pillars, the inputs and the alternative choices, as here, so that the reader can disagree with a weight rather than with an opaque score.
Question. How does all this become a report a portfolio manager will read in five minutes?
Page one carries the title, which states the view (“Acierna: priced for a transition it has not yet funded; initiate at Hold”), the rating, the target price, the three to five key points and a table of forecasts and valuation metrics. Léa writes the first key point on carbon, because it drives the valuation: the free allocation is worth almost as much as 2026 EBITDA, and the target price falls from EUR 18.56 carbon-blind to EUR 10.69 once carbon and the plan are modelled. The second point quantifies the debate with the market (74% implied, 60% assumed). The third gives the catalysts: the Council position on the ETS review, phase 2 funding, and 2027 price data on CBAM pass-through. The fourth covers the risks to the rating in both directions: a slower phase-out or full pass-through would push the stock to Buy; a gas price spike or a grant refusal would push it to Sell.
The ESG section is not a list of scores from vendors. It contains the carbon bridge from reported to carbon-adjusted EBITDA, the scorecard with its weights, the data sources and their dates, and a paragraph on what the analysis does not cover: scope 3 emissions from iron ore and coal mining, local air quality, and the social consequences of closing a blast furnace, which can delay the plan politically and are material to delivery. Research compliance checks the wording: the report says “transition capex”, not “green capex”, and does not call the plan “Paris-aligned”, since no method in the report supports that claim.
The credit initiation is published the same day with its own page one: Acierna remains BBB− in Meridian’s view, with negative pressure in 2029-2032, underweight on the long end, and a list of balance-sheet actions that would remove the pressure.
| Figure | Value | Block |
|---|---|---|
| Scope 1 emissions 2026 | 13.54 Mt | c13-carbon-bill |
| Net carbon cost 2026; 2030; 2034 | EUR 194.0m; 843.6m; 1,895.6m | c13-carbon-bill |
| Value of free allocation 2026 | EUR 889.2m | c13-carbon-bill |
| Reported EBITDA 2026; full carbon cost EBITDA | EUR 1,047.4m; 158.2m | c13-pass-through |
| Share of 2030 bill recovered through prices | 86.7% | c13-pass-through |
| Carbon-adjusted EBITDA 2030; 2034 (no plan) | EUR 887.9m; 604.6m | c13-pass-through |
| Carbon-adjusted EBITDA 2030; 2034 (plan) | EUR 1,145.4m; 1,314.6m | c13-transition-plan |
| Emissions cut by 2034; transition capex share | 48.0%; 66.7% | c13-transition-plan |
| Abatement cost of phase 1 | EUR 83.9/t CO2 | c13-transition-plan |
| Target price carbon-blind; no plan; plan | EUR 18.56; 5.91; 13.88 | c13-valuation |
| Weighted target price; upside; rating | EUR 10.69; −9.4%; Hold | c13-valuation |
| Market-implied probability of delivery | 73.9% | c13-valuation |
| Peak leverage (2032); lowest FFO/debt | 3.17x; 23.4% | c13-credit-path |
| Bond price change on downgrade | −4.4 points | c13-credit-path |
| Transition score (primary route; company) | 63.0; 70.8 | c13-scorecard |
Research holds no securities and lends no money, so it creates no credit or market risk of its own. Its capital is operational-risk capital. Under CRR3 (Regulation (EU) 2024/1623), applicable since 1 January 2025, operational risk is computed with the standardised approach: the business indicator (interest, services and financial components; research payments and commissions sit in the services component) is multiplied by marginal coefficients of 12% up to EUR 1bn, 15% from EUR 1bn to 30bn and 18% above, and the EU sets the internal loss multiplier to 1. Meridian’s business indicator lies in the middle bucket, so each euro of revenue attracts 15% of capital requirement, or 12.5 × 15% = 1.875 euros of RWA. Attributing this to a team is a management convention, not a regulatory rule, but it is the right marginal logic.
The operational risks of research are conduct risks: publishing inside information, a conflict not disclosed, a misleading ESG claim, a model error in a widely read report. They are rare and can be expensive; they are reflected in the bank’s loss history and in Pillar 2, not in a specific charge for the team.
Meridian’s management accounts treat the metals and mining research franchise as a unit with two revenue lines and an explicit cost base. Because research costs are fixed people and data costs, the 60% cost-income convention used for fee businesses in this course is replaced by the actual cost base; the convention is shown for comparison. The case assumptions are: research payments from clients for metals and mining EUR 1.6m; cash equity commissions on the sector’s stocks attributed to research under Meridian’s attribution key EUR 1.4m; team cost (lead analyst, associate, junior) EUR 1.65m; share of an ESG specialist EUR 0.25m; financial data EUR 0.20m; ESG and carbon data EUR 0.15m; publishing, compliance and infrastructure allocation EUR 0.40m. No investment banking revenue is attributed to research: the independence rules forbid linking research budgets or analyst pay to deals.
INTERACTIVE LAB — RAROC of Meridian’s metals and mining research franchise
Compute the revenues, costs, operational-risk capital and RAROC of the research franchise, and the cost of its ESG and climate layer.
| Parameter you can move | Range | Default |
|---|---|---|
| Research payments from clients | 0 to 5 EUR m | 1.6 EUR m |
| Attributed trading commissions | 0 to 5 EUR m | 1.4 EUR m |
| Analyst team cost | 0.5 to 4 EUR m | 1.65 EUR m |
| ESG specialist share | 0 to 1 EUR m | 0.25 EUR m |
| Financial data | 0 to 1 EUR m | 0.2 EUR m |
| ESG and carbon data | 0 to 1 EUR m | 0.15 EUR m |
| Publishing, compliance and infrastructure | 0 to 1 EUR m | 0.4 EUR m |
| Marginal business-indicator coefficient | 12 / 15 / 18 | 15 % |
| Result at default settings | Value |
|---|---|
| Revenues | 3.00 EUR m |
| Costs | 2.65 EUR m |
| Cost-income ratio | 88.3 % |
| Cost of the ESG layer | 0.40 EUR m |
| Operational-risk RWA | 5.625 EUR m |
| Allocated capital (13%) | 0.731 EUR m |
| Profit after tax | 0.2625 EUR m |
| RAROC | 35.9 % |
| RAROC without the ESG layer, same revenues | 76.9 % |
| RAROC with the 60% cost-income convention | 123.1 % |
| Extra revenue needed for the ESG layer to earn the 12% hurdle | 0.416 EUR m |
Try it. Cut research payments to EUR 1.0m, as after a poor broker review: the franchise loses money. That is the commercial pressure behind every research budget negotiation.
| Line | Value |
|---|---|
| Revenues (payments 1.6 + commissions 1.4) | EUR 3.00m |
| Costs (of which ESG layer 0.40) | EUR 2.65m |
| Expected loss | none |
| Operational-risk RWA | EUR 5.625m |
| Allocated capital (13%) | EUR 0.731m |
| Profit after tax | EUR 0.263m |
| RAROC | 35.9% |
Reading. The RAROC of 35.9% looks comfortable only because operational-risk capital is tiny; the cost-income ratio of 88.3% tells the real story of a thin margin. With the course’s 60% convention the RAROC would be 123.1%, which shows why research must be managed on its actual costs. The ESG and climate layer costs EUR 0.40m a year; without it the franchise would earn 76.9% on the same revenues. To pay for itself at the 12% hurdle the layer must bring EUR 0.416m of extra revenue a year, for example by moving two or three clients up one tier in the broker review. Meridian’s head of research argues that the carbon module is what makes the Acierna initiation worth reading, but cannot prove the revenue attribution, which is the honest state of research economics.
The indirect value lies elsewhere in the bank. Credit research supports Meridian’s market making in Acierna bonds (see C08), and equity research supports the cash equities franchise. The bank cannot, however, count on research to win the mandate if Acierna later issues a transition bond: that link is exactly what the independence rules break.
Research uses no balance sheet. Research payments are invoiced quarterly or annually and create small receivables; the FTP charge on them is negligible. There is no HQLA requirement and no liquidity risk worth modelling.
Research payments are revenue from contracts with customers under IFRS 15. An annual research subscription is a service delivered over time and is recognised over the period; a payment for a specific bespoke piece of work is recognised when it is delivered. Under joint payment after the Listing Act, the commission paid by a client covers both execution and research, and the broker has to decide whether these are separate performance obligations and how to allocate the price, which most treat as commission revenue recognised at trade date with internal attribution to research. Data licences are expensed as incurred; cloud-based data services generally do not meet the definition of an intangible asset that the bank controls, as the IFRS Interpretations Committee explained in its agenda decisions on cloud computing arrangements. Analysts’ compensation is a staff cost, with deferral rules for material risk-takers under CRD remuneration requirements where relevant.
Green Asset Ratio and Pillar 3. Research creates no exposure and therefore enters neither the GAR nor the Pillar 3 ESG templates. Financed and facilitated emissions. PCAF’s standard for facilitated emissions (Part B, 2023) covers underwriting of bonds, equity and syndicated loans; research is outside it. Acierna’s emissions enter Meridian’s metrics only through its loans, its underwriting and its trading positions (see E07). Own reporting and reputation. Research matters to the bank’s own sustainability statement in a different way: published research that contradicts the bank’s stated climate policy, or that relies on ESG claims the bank cannot support, is a reputational and conduct issue that the bank’s greenwashing controls must cover (see E10). Meridian therefore requires every sector to use the house carbon price deck and the house definitions of “transition” and “green”, which the ESG research team maintains.
| Risk | Control | Owner |
|---|---|---|
| Banker pressure on rating or coverage | Information barriers, minuted coverage decisions, no IB-linked pay | Head of research, compliance |
| Publishing inside information | Watch and restricted lists, wall-crossing log | Control room |
| Undisclosed conflict | Disclosure engine, pre-publication review | Research compliance |
| Model or formula error | Second-person model review, version control | Lead analyst, research quality |
| Vendor estimates presented as reported data | Source hierarchy: registry, filings, then vendor | Data team, ESG research |
| Inconsistent carbon price across sectors | Single house carbon price deck | ESG and climate research |
| Greenwashing in report wording | House definitions, ESG wording checklist | Research compliance |
| Selective disclosure in management meetings | Meeting rules, compliance chaperone when needed | Analyst, compliance |
| ESG score falling under the ESG Ratings Regulation | Legal scoping, no stand-alone score products | Legal, head of research |
| Joint payment hiding research value | Annual quality assessment, client reporting | Research management, finance |
Red flags in an ESG or climate section.
Research payment. MiFID II unbundling has applied since 3 January 2018. The UK re-allowed joint payment from August 2024. In the EU the Listing Act (Directive (EU) 2024/2811) removed the EUR 1bn threshold for joint payment, with transposition due by 5 June 2026 and application from 6 June 2026; Delegated Directive (EU) 2026/374 was published in the Official Journal on 2 June 2026, and requires an annual quality assessment of research whatever the payment method. Not all member states had transposed on time. Whether asset managers return to bundled payment at scale, and whether mid-cap coverage recovers, are open questions at the end of 2026.
ESG ratings. The ESG Ratings Regulation has applied since 2 July 2026; existing providers had to notify ESMA by 2 August 2026 and apply for authorisation or recognition by 2 November 2026 to keep operating during the transition. ESMA’s register of authorised providers is the reference for analysts quoting third-party ratings.
Carbon and steel policy. The December 2026 EUA contract traded around EUR 90 per tonne in early November 2026. The Commission’s ETS review proposal of 17 July 2026 would slow the CBAM phase-in; the Council aims for a position on 11 December 2026, and analysts treat the July text as a negotiating baseline. CBAM certificates go on sale on 1 February 2027. The EU’s permanent steel trade measure applies since 1 July 2026, with quotas cut by 47% on average and a 50% out-of-quota duty. CSRD scope was narrowed by the Omnibus I Directive in force since 18 March 2026.
Industry. worldsteel’s 2025 indicators give 1.92 t CO2 per tonne of crude steel globally for 2024. European producers have scaled back several hydrogen-based plans since 2024: ArcelorMittal, for instance, confirmed in February 2026 a EUR 1.3bn, 2 Mt electric arc furnace at Dunkirk due in 2029, after setting aside its earlier direct-reduction project for the site. Analysts now value transition plans as options with execution risk rather than as committed trajectories.
Credit ratings and ESG. Agencies integrate ESG where material and publish relevance or impact scores, except S&P, which dropped separate ESG indicators in 2023. The ESG content of credit research is increasingly expressed as rating-threshold analysis through capex cycles, as in the case, rather than as scores.
Open debates. How much of CBAM will pass through to EU steel prices, and how fast; whether ESG and carbon premia in returns exist once data quality is controlled; whether issuer-sponsored research under the ESMA code will be trusted by investors; and how far bank research ESG scores fall within the ESG Ratings Regulation.
INTERACTIVE LAB — Research and ESG integration: check your understanding
Test the mechanics and the critical points of the module.
Q1. Under the Listing Act, from 6 June 2026 an EU asset manager may pay for research jointly with execution:
Answer: B. Directive (EU) 2024/2811 removed the EUR 1bn threshold and added conditions on methodology, information and quality assessment.
Q2. Before publication, the issuer may review:
Answer: C. Independence rules allow a factual check but not disclosure of the rating, target price or valuation.
Q3. Acierna’s 2026 net carbon cost of EUR 194m is computed as:
Answer: C. Shortfall after applying the 2026 CBAM factor to the base allocation, times the price.
Q4. The CBAM factor in the law in force for 2030 is:
Answer: C. Article 10a(1a) of the ETS Directive: 51.5% in 2030; the July 2026 proposal would raise it to 59%.
Q5. Why can a producer recover more than its own carbon cost on domestic sales?
Answer: B. Pass-through is market-driven: the uplift follows the marginal supplier’s cost.
Q6. Subtracting emissions × EUA price from reported EBITDA:
Answer: B. Reported EBITDA already bears the net purchase cost; start from pre-carbon EBITDA instead.
Q7. The share price of EUR 11.80 lies between the no-plan (5.91) and plan (13.88) values. The implied probability of delivery is about:
Answer: B. (11.80 − 5.91) / (13.88 − 5.91) = 73.9%.
Q8. Adding a climate premium to the WACC is justified when:
Answer: C. Otherwise the same risk is counted twice.
Q9. A long-short ESG strategy has a true premium of 2% and volatility of 8%. Years needed for an expected t-statistic of 2:
Answer: C. (2 × 8 / 2)^2 = 64 years.
Q10. Pástor, Stambaugh and Taylor explain the strong realised returns of green stocks in the 2010s mainly by:
Answer: B. Realised returns rose as concerns and flows shifted; expected returns of green assets should be lower in equilibrium.
Q11. Why does the credit analyst recommend underweight on Acierna’s 2031-2033 bonds while the plan helps equity value?
Answer: B. Bondholders bear the capex peak; the carbon savings come largely after their maturity.
Q12. Comparing Acierna’s company-average intensity with a sector pathway:
Answer: B. The company average scores the maximum because 30% of output is scrap-based, not because of the plan.
Q13. Research revenue and capital at Meridian: which statement is correct?
Answer: C. Independence rules forbid linking research to deals; capital is operational risk under CRR3.
Q14. Which emissions framework covers research activity?
Answer: C. Research is outside PCAF’s financed and facilitated emissions standards and outside the GAR.
A producer emits 5.0 Mt of CO2 a year, has a base free allocation of 4.2 Mt and sells 3.0 Mt in the EU. In 2030 (CBAM factor 51.5%) the EUA price is EUR 100. The marginal import has 1.90 t CO2 per tonne, the benchmark-equivalent allocation is 1.40 t per tonne, and 70% of the import carbon cost reaches EU prices. Compute the net carbon cost, the uplift and the net effect on EBITDA.
Solution. Free allocation: 4.2 × 0.515 = 2.163 Mt. Shortfall: 5.0 − 2.163 = 2.837 Mt. Net cost: 2.837 × 100 = EUR 283.7m. Import carbon cost per tonne: (1.90 − 1.40 × 0.515) × 100 = (1.90 − 0.721) × 100 = EUR 117.9. Uplift: 70% × 117.9 = EUR 82.53 per tonne, or 82.53 × 3.0 = EUR 247.6m. Net effect: 247.6 − 283.7 = −EUR 36.1m. The producer recovers 87% of its bill; the remainder comes from its exports and from emitting more per tonne sold domestically than the uplift compensates.
A stock trades at EUR 22. The analyst values it at EUR 30 if the transition plan is delivered and EUR 12 if not, and assigns a 40% probability to delivery. The rating system is Buy above +15%, Sell below −15%.
Solution. Weighted target price: 0.4 × 30 + 0.6 × 12 = EUR 19.20. Upside: 19.20 / 22 − 1 = −12.7%: Hold, close to Sell. Implied probability: (22 − 12) / (30 − 12) = 55.6%. The analyst is 15.6 points more sceptical than the market and should say why, for example by pointing to the weakest pillar of the scorecard.
A research team earns EUR 5.0m (payments and attributed commissions) and costs EUR 4.2m. The marginal business-indicator coefficient is 15%, capital 13% of RWA, tax 25%.
Solution. Operational-risk RWA: 12.5 × 15% × 5.0 = EUR 9.375m. Capital: 13% × 9.375 = EUR 1.219m. Profit after tax: (5.0 − 4.2) × 0.75 = EUR 0.60m. RAROC: 0.60 / 1.219 = 49.2%. The cost-income ratio of 84% is the better indicator of fragility: a 16% fall in revenues wipes out the profit.
1. How would you integrate carbon into the valuation of an EU steelmaker? I would start from EBITDA before carbon, then deduct each year’s net carbon cost: emissions minus free allocation after the CBAM factor, times a house price deck. I would add the EU price uplift from CBAM on domestic shipments, with an explicit pass-through share, and nothing on exports. I would keep the WACC unchanged so as not to double-count. Finally I would value scenarios with and without the transition plan and weight them, showing the probability the share price implies.
2. What changed for research payment in 2026? The Listing Act removed the EUR 1bn market capitalisation threshold for joint payment, applicable from 6 June 2026. Asset managers may now pay for research through execution commissions for any issuer, provided they agree a remuneration methodology, inform clients and assess research quality annually. The UK had re-allowed bundling in 2024. The effect on research budgets and coverage is still being observed.
3. Does ESG investing outperform? The honest answer is that the evidence is mixed and period-dependent. Realised outperformance in the 2010s is largely explained by unexpected shifts in climate concern, while equilibrium theory predicts lower expected returns for green assets. The carbon premium literature is sensitive to whether emissions are disclosed or estimated. And most ESG datasets are too short to give statistical power for premia of a few per cent. I would trust ESG more as a risk and cash-flow input than as a source of alpha.
4. Why might the equity and credit analysts disagree on the same company? They have different horizons and pay-offs. The equity holder captures the long-term benefit of a transition plan, including the terminal value. The bondholder gains little beyond the coupon and bears the risk that capex pushes leverage over a rating threshold before maturity. In the Acierna case, the same model gives a plan that adds equity value but creates rating pressure in 2029-2032.
5. What is the main weakness of ESG ratings for an analyst? They disagree with each other much more than credit ratings do, because providers measure different things with different weights. Scores can also be revised retrospectively and often rely on estimated data when companies do not disclose. I would use them as a starting point and a client-facing reference, and rebuild the material issues myself from filings and registries. Since July 2026, providers in the EU are supervised by ESMA, which should improve transparency of methods but not make scores converge.
6. How do you make sure your research is independent? The coverage decision is minuted and not driven by bankers. I do not take part in pitches, my pay is not linked to investment banking revenue, and the issuer sees only factual sections. Any wall-crossing goes through the control room. I document assumptions and scenarios so that compliance and readers can trace every figure.
7. What is pass-through and why is it so important in steel? Pass-through is the share of a cost increase recovered through higher prices. In steel it is set by the market, by the cost of the marginal supplier, which under CBAM is increasingly the carbon cost of imports. In the case, moving from 80% to 100% changes 2034 EBITDA from about EUR 605m to about EUR 980m. Any carbon-adjusted valuation that does not state its pass-through assumption is incomplete.
8. A client asks you to rank steelmakers by “greenness”. What do you answer? I would ask what decision the ranking supports. Average intensity rewards producers with high scrap shares, which says more about their asset base than about their transition. I would rank on primary-route intensity trajectory, transition capex as a share of total capex, the funding and delivery risk of the plan, and governance. I would show the weights and alternatives, and avoid calling any of them “green” without a stated standard.
Regulation and official texts
Market and policy sources consulted (November 2026)
Academic literature