Heavy
Lifting.

This piece is co-authored with Shivaram Rajgopal and available on The Googly.

The first published picture of the platypus, from George Shaw’s Naturalist’s Miscellany, 1799. Shaw confessed it was “impossible not to entertain some doubts as to the genuine nature of the animal.”

Frederick Polydore Nodder, in George Shaw, The Naturalist’s Miscellany, vol. 10, 1799. Biodiversity Heritage Library. Public domain, via Wikimedia Commons.

At the valuations reportedly under discussion for its public offering, between $1.5 trillion and $2 trillion, Anthropic must grow its profits by somewhere between 27 and 76 percent a year for five years to reward the investors who buy in. Where it falls in that range depends on the price and on a question its registration statement, filed on October 26, does not answer directly. As far as we can find, no company of Anthropic’s size has ever grown faster than 35 percent a year for five years.

Almost everything written about the S-1 so far has been about revenue: $16.3 billion in the first half of 2026, growth of 1,300 percent, an annualized run-rate reported at $65 billion, and a first-half loss of $80.7 billion that is mostly an accounting remeasurement of convertible notes. Much less has been said about how much of that revenue the company keeps. A few analysts, Tomasz Tunguz and CJ Gustafson among them, have dug a gross margin of 85 percent out of a footnote. Our question is what that margin means, because the answer decides which kind of business Anthropic is, and therefore how much growth its price demands.

A Creature That Will Not Classify

When the first platypus specimen reached Britain in 1799, naturalists suspected a hoax: a duck’s bill sewn onto a mammal. The animal lays eggs, nurses its young and has a bill. It fit none of the existing categories. Charles Darwin, travelling inland from Sydney in January 1836, watched platypuses in a creek and, set beside the strange marsupials around them, imagined a skeptic exclaiming, “Surely two distinct Creators must have been at work.” Minutes later he found an antlion almost identical to its European cousin and took the thought back. Creatures that look unclassifiable sometimes turn out to belong to a known family after all.

Charles Darwin in 1840, four years after he watched platypuses in a creek west of Sydney and briefly wondered whether “two distinct Creators must have been at work.”

George Richmond, Charles Darwin, watercolor, 1840. English Heritage, Down House. Public domain, via Wikimedia Commons.

Frontier AI laboratories are a new creature too. Investors are trying to decide which known species they belong to, and there are three candidates. We call them the spider, the elephant and the shark. Anthropic’s S-1 is the first full specimen any laboratory has submitted for public examination, and it contains evidence for all three, in three different measures of profit.

Three Measures of Profit

In 2025, Anthropic stopped reporting cost of revenue and gross profit on the face of its income statement. It now groups expenses by type: compute and infrastructure, personnel, sales and marketing, and administration. A note to the financial statements still gives the older breakdown by function, including cost of revenue and research and development. From these two presentations, three measures of profit can be calculated, as Table 1 shows. The S-1 states none of them.

All three are correct. They differ in what they count as the cost of the product. The gross margin counts only the cost of serving paying customers. The margin after all compute counts every unit of computing power the company uses, including what it spends training new models, serving free users and running Claude for its own staff. The margin after R&D counts the cost of serving customers plus all model development. Which one describes the business depends on a single question: is training new models an investment, or a recurring cost of what Anthropic sells?

Two features of the filing flatter the gross margin. The compute used to serve free users is recorded in sales and marketing rather than cost of revenue, so it never touches the gross margin at all. And the gross margin itself has moved sharply: 12 percent in the first half of 2025, 40 percent for 2025 as a whole and 85 percent in the first half of 2026. The S-1 does not explain the jump.

The Spider

A spider spins a web and catches prey in it again and again at almost no extra cost. That is the economics of software. Building the product is expensive; selling one more subscription is cheap. Accountants treat the building as research and development, an investment recorded below gross profit, and mature software companies keep 75 to 85 cents of each dollar of sales as gross profit. Investors pay about $7 for each dollar of annual gross profit of the median public software company, and far more for fast growers: about $54 for Palantir, whose revenue rose 93 percent last quarter.

Maria Sibylla Merian’s plate of Suriname spiders, 1705. The ones above catch their prey in webs; the one below seizes a hummingbird without one, a scene naturalists dismissed as fancy until Henry Walter Bates watched a bird-eating spider kill two finches in the Amazon. Our spider needs its web.

Maria Sibylla Merian, engraved by Pieter Sluyter, Metamorphosis insectorum Surinamensium, plate 18, 1705. Public domain, via Wikimedia Commons.

On its 85 percent gross margin, Anthropic looks like a spider, with gross profit of about $55 billion a year at its current run-rate.

One caution about the metaphor. Even spiders rebuild their webs, many of them every night. Software companies rebuild theirs every few years. How often the web must be rebuilt turns out to be the whole question.

The Elephant

An elephant must eat constantly to sustain a body it cannot shrink. That is the economics of capacity businesses: cloud providers, telecommunications networks, data centers. They buy enormous capacity in advance, and their profit depends on filling it. Investors watch utilization above all, because overbuilt capacity is how such businesses fail. The cost that matters is the cost of all the capacity, not only the share used for paying customers.

Jumbo, the London Zoo’s African elephant, on the cover of sheet music published in Boston in 1882, the year P. T. Barnum bought him for £2,000 despite a national outcry in Britain. An elephant is worth what it draws and costs what it eats.

“Jumbo,” sheet-music cover, Oliver Ditson & Co., Boston, 1882; lithograph by J. H. Bufford’s Sons. Library of Congress. Public domain, via Wikimedia Commons.

The margin after all compute, 45 percent, is the elephant’s measure, and it gives Anthropic about $29 billion a year. A mature elephant such as Equinix, which owns data centers and grows about 12 percent a year, is priced at about $22 for each dollar of profit after the cost of its capacity. That is a generous price, because Equinix’s buildings last for decades. CoreWeave, which rents out AI computing capacity, is priced at about 80 times its much thinner margin, measured the same way, because investors are paying for roughly $100 billion of signed contracts: capacity they expect it to fill.

The Shark

Many sharks must keep swimming to breathe. That is the economics of a content treadmill. Netflix’s shows lose their audience, so it must keep making new ones, and it accounts for them accordingly: it capitalizes what it spends producing and licensing content and amortizes that cost, front-loaded, over the expected period of viewing, inside its cost of revenues. Film studios do the same under a dedicated accounting standard. The cost of making the product is treated as part of the cost of selling it. Netflix keeps about 53 cents of each dollar after that cost, and investors pay about $11 for each dollar it keeps.

Brook Watson, aged 14, lost his right leg to a shark in Havana harbor in 1749. He went on to chair Lloyd’s of London and to serve as Lord Mayor. Investors buying a shark at a spider’s price should be so lucky.

John Singleton Copley, Watson and the Shark, 1778. National Gallery of Art, Washington, Ferdinand Lammot Belin Fund. Public domain, via Wikimedia Commons.

Anthropic’s models also go stale, and the S-1 says how fast. Model versions released in the six months to September 30 accounted for more than 80 percent of third-quarter revenue. Four-fifths of what Anthropic sells is less than six months old. Staying at the frontier means training new models continuously, and the old ones fade quickly once the new ones arrive. If training is a cost of the product, the right measure is the margin after R&D, 26 percent, or about $17 billion a year. That figure is conservative, because R&D also includes researchers’ pay and other costs that are not training.

The analogy has a limit, and a well-informed reader will press on it. When a Netflix show goes stale, little of it carries over to the next one. When a model goes stale, much does: the methods, the data and the lessons of training it. The S-1 describes each Claude model helping to build the next, a dynamic it calls “closing the loop.” That is the spider’s best argument: training builds lasting capability, even though each model earns money only briefly. But if what carries over made the next model cheaper to build, the bill would be shrinking. It is growing. Research and development spending rose from $1.9 billion in the first half of 2025 to $9.7 billion in the first half of 2026. The knowledge compounds, and so does the cost, while the revenue from each model still fades within months.

Anthropic does not account for training the way Netflix accounts for shows. It expenses training immediately as research and development and reports that capitalized software costs were “not significant.” That is a judgment, not a requirement. Accounting advisers have argued that development-stage training costs can be capitalized under the standard for internal-use software, which would put them on the balance sheet and amortize them, as cost of revenue, over each model’s life. A company whose training runs sometimes fail, as Anthropic’s S-1 says its runs have, can reasonably treat them as research. But the choice keeps the cost of training out of the gross margin entirely.

Netflix’s figure involves judgment too. It spreads each show’s cost over the period it expects the show to be watched: more than 90 percent within four years, most of it early. If shows lose their audience faster than that, Netflix books their cost too slowly and its profit is overstated. But a longer or shorter schedule only changes when costs are booked, not how much, and Netflix spends roughly the same on content each year, so the gap stays small. In 2025 it charged $16.4 billion of content amortization against $17.7 billion of cash content spending, a difference of about 3 points of margin. For Anthropic, timing cuts the other way. It writes off training immediately, and because that spending is growing fast, the immediate write-off understates its profit compared with spreading the cost over even a single year. If the shark comparison is tilted, it is tilted against Anthropic.

Turning the Question Around

There is no need to guess what Anthropic is worth. Its bankers will set a price. The useful question is what that price requires.

Suppose an investor buys at the offering and wants the long-run return of the stock market, about 10 percent a year. Suppose too that within five years Anthropic matures into an ordinary member of its species and is priced like one, keeping today’s margins. How fast must its profits grow in the meantime? The answer depends on the species, because each counts profit differently and each is priced differently when mature, as Table 2 shows. Analysts often compare companies by dividing market value by annual revenue, but that shortcut assumes similar margins, and margins are the whole question here. So we compare price with profit, measured for each comparable the same way as for Anthropic.

Anthropic may also be all three at once: a spider in how it sells, an elephant in how it buys compute and a shark in how often it must retrain. So we add a composite, valued as one-third of each.

Table 3 gives the growth each valuation requires, over five years and over ten.

At $2 trillion, the top of the reported range, Anthropic must grow its profits by 36 percent a year for five years as an elephant, 49 percent as the composite, 51 percent as a spider and 76 percent as a shark. Spread over ten years, the requirement falls to between 22 and 39 percent a year, but it must then be sustained twice as long. Even at $1 trillion, half that price, the shark needs 51 percent a year for five years.

How Extraordinary That Would Be

Size is what makes these numbers hard. At a $65 billion run-rate, Anthropic already ranks among the largest companies in the world by revenue, and large companies rarely grow fast. Table 4 compares the required rates with the fastest growth companies of that size have achieved.

The record is short. Across all industries from 1950 to 2015, no company with more than $50 billion in sales grew faster than 35 percent a year over five years. More recently, Amazon managed 29 percent from 2015 to 2020 and Alphabet 23 percent from 2016 to 2021. Nvidia grew 88 percent a year from fiscal 2024 to 2026, the fastest run at this scale we know of, but for two years, not five.

Measured against that record, every reading of a $2 trillion price asks for something new. The elephant asks for growth just beyond the best five years on record. The spider and the composite ask for about 1.7 times Amazon’s best, and the shark for more than twice it. Over ten years the gap is similar: the best decade at this size we know of is Alphabet’s, at about 19 percent a year from 2013 to 2023, below what any species requires at $2 trillion.

Anthropic does have a head start. It is growing at about 1,300 percent a year, and the first year counts heavily. If it tripled its revenue next year and kept its margins, the spider would need 27 percent a year for the following four years, close to the record, while the shark would still need 54 percent, nearly twice the record.

The species matters most at the extremes. The spider and the elephant can reach a $2 trillion price with growth that is unprecedented but not wildly so, especially after a strong first year. The shark cannot, unless Anthropic sustains growth for years at more than twice any pace yet seen at its size.

How Anthropic Sees Itself

The S-1 is clear about which measure management uses. It says Anthropic deploys compute “as a single, shared resource across all workloads” and assesses it “on a total basis rather than by individual workload.” Its account of its own financial strength rests on total compute against revenue: in the second quarter of 2026, revenue grew 1,368 percent from a year earlier while compute grew 300 percent, and compute fell from 251 percent of revenue in the first quarter of 2025 to 48 percent in the second quarter of 2026. That is the elephant’s measure. The company also says that serving customers and developing models each feed the other, which is why it stopped presenting costs by function. The gross margin of 85 percent appears only in a note, without comment.

So Anthropic manages itself like an elephant, describes its product like a spider and depends, by its own account, on staying at the frontier like a shark. Management’s own measure is the one that demands the least growth. The measure that matches its own account of why customers buy demands the most.

What the Evidence Says

The S-1 says more about the classification than it first appears. The commercial life of a model is the question that separates the spider from the shark: if models stay useful for years, training looks like a spider’s investment; if they go stale in months, it looks like a shark’s production cost. The 80 percent figure answers it about as clearly as a filing can. Most of Anthropic’s revenue comes from models that did not exist half a year earlier.

The S-1 also shows how little of Anthropic’s compute serves paying customers. Cost of revenue was $2.4 billion in the first half of 2026, against $8.9 billion of total compute and infrastructure, so at most about 27 percent of the compute went to serving paying customers. The rest went to research, building new models, free users and Anthropic’s own staff. That is the proportion of a laboratory that sells its output, not of a software company that maintains a product.

What the S-1 does not give is the piece that would make the shark’s accounts precise: how much of the remaining compute went to training new models, as opposed to research and internal use, and how much revenue each model earns over its life compared with what it cost to train. Anthropic has both; the S-1 describes the capacity-management system that allocates compute across workloads. A third figure would sharpen the picture: the cost of the compute used to serve free users, now hidden in sales and marketing.

A Shark at a Spider’s Price

Naturalists eventually stopped trying to fit the platypus into an existing category. They placed it, with the echidnas, in an order of its own, the monotremes: mammals that lay eggs. Anthropic may need the same treatment, since its training is part investment and part treadmill, and a measure built for it rather than borrowed. The obvious candidate is the spider’s gross margin with one change: a charge for training, spread over each model’s commercial life, as Netflix spreads the cost of a show over its viewing. That measure needs the two numbers the S-1 leaves out.

By the time John Gould painted it, the platypus had an order of its own.

John Gould and Henry Constantine Richter, Ornithorhynchus anatinus, from Gould’s The Mammals of Australia, 1845–63. Biodiversity Heritage Library / Smithsonian Libraries and Archives. Public domain.

Until they are disclosed, the filing’s own evidence points one way. Anthropic’s products go stale within months, most of its compute goes to making the next one, and its margin after paying for that is a fraction of its gross margin. On that evidence the creature looks more like a shark than a spider, and the shark is the species whose $2 trillion price asks for growth more than twice as fast as any company of its size has sustained. An investor buying at that price is betting that the shark becomes a spider: that training will one day cost less relative to what each model earns. The S-1 gives good reason to ask how, and no numbers that show when.

A Note on Method

We wrote this as a hybrid team of two authors and AI models. Each of us worked with Claude, Anthropic’s own model, a fact readers should weigh: both of us used Claude Opus 5.5. Google’s Gemini critiqued the plan for this series and the draft, and its objection to the shark analogy is answered above. The models helped read the 235,000-word filing and the comparables’ reports. We checked every figure taken from the S-1 against the filing. Market prices, comparables’ margins, and historical growth rates come from the sources noted under the tables. The required growth rates assume constant margins, a 10 percent annual return, and a mature price reached after five or ten years; changing any of these changes the numbers, but not their ranking across species.

Appendix: How This Piece Was Made

This is the first of four weekly pieces on Anthropic’s S-1, each built around an animal. This one classifies the specimen. “Feeding the Elephant” describes how it feeds: a shark living in an elephant’s body, training new models on computing capacity bought years in advance from companies that are also its investors, distributors, and competitors. “Can a Shark Stop Swimming?” asks whether Anthropic’s public-benefit purpose can bind it, and what pausing would cost. “The Paper Tiger” explains why the $80.7 billion loss in the headlines is not the business. Each piece will carry an appendix like this one, because we think the way a piece is produced is part of what a reader should be able to judge.

How the Piece Evolved

It began with Shiva. In the days after the S-1 was filed, he ran it through analytical tools he has built for boards and investors, prompt libraries for governance red-teaming, strategy, and earnings quality, using Claude to apply them to the 235,000-word filing. The result was seven analyses, from the governance of Anthropic’s trust to the cost of a safety pause. He also made a point that turned out to be the seed of this piece. Anthropic had stopped showing cost of goods sold on the face of its income statement.

Bob and Claude built the first draft on that point in a day. Claude found that a note to the financial statements still gives the old breakdown, and from it and the income statement derived the three measures of profit in Table 1. Bob proposed an animal analogy to make the measures intelligible; Claude chose the three species and matched each to a measure, with the platypus as the specimen that will not classify. That draft asked what Anthropic would be worth as each animal, by applying each species’ typical valuation multiple, and concluded that it was priced as a spider but looked more like a shark. It went to Shiva with eight questions, the last of which was whether the animals were too cute.

Shiva kept the animals and changed the method. Valuing Anthropic at other companies’ multiples, as our own draft conceded, depends heavily on which comparables one picks and mixes different bases. Shiva took the price as given and asked what it requires: how fast profit must grow, on each species’ measure, for an investor buying at the offering to earn a market return, and how that compares with what companies of Anthropic’s size have ever achieved. That turned a valuation exercise into a test against the historical record, and produced Tables 2 to 4 and the opening of the piece. He also corrected our account of Darwin, who had imagined a skeptic’s exclamation rather than making it himself.

We then brought the two versions together. Shiva’s revision became the base, and material from our earlier draft was restored: the S-1’s evidence that more than 80 percent of revenue comes from models less than six months old, the share of compute that serves paying customers, and the caveat, prompted by a critique from Google’s Gemini, that much carries over from one model to the next even as each goes stale. Bob read the result as a general reader would and stopped wherever he could not follow, which removed one confusing passage altogether. Finally, Bob found and downloaded the public-domain images, using an approach from his earlier pieces, Claude proposed the subjects and wrote the captions, and Shiva audited the open questions, as described below.

Who Contributed What

The two of us contributed differently. Shiva supplied the analytical substrate, the accounting insight at the start, the change of method in the middle and the audit at the end. Bob supplied the framing, orchestrated the work across people and models, and acted as the reader who refuses to nod along. Claude, working alone, read and searched the filing, ran the calculations and searched the specialist literature; working with each of us, it drafted, restructured, and checked. Gemini served as an outside critic. Several of the best ideas belong to no single participant but to the exchanges between them. The two of us take responsibility for every claim.

How We Checked the Numbers

When the draft was finished, four figures still carried notes saying we were not yet sure of them. Before publication Shiva resolved each one, working with Claude, and recorded what he checked, how he calculated it and what changed. Two of the four checks changed the text and two confirmed it.

Palantir’s price: confirmed. We wrote that investors pay about $54 for each dollar of Palantir’s gross profit, while a market recap had put the figure near 90. Both are right on different bases. Palantir’s enterprise value of about $464 billion is roughly 54 times its expected gross profit for the coming year, and about 89 times its gross profit over the past year. We use the forward figure because it matches how we price the median software company in Table 2; the recap was quoting the trailing one.

CoreWeave’s price: corrected from about 90 times to about 80. To compare CoreWeave fairly with Equinix, its profit has to be measured the same way: revenue minus the full cost of its capacity, including the depreciation of its chips, which CoreWeave reports outside its cost of revenue. On that basis its margin in the second quarter of 2026 was about 7 percent, and its enterprise value of about $78 billion is roughly 80 times the resulting profit on its 2026 revenue guidance. The figure is very sensitive to the measure chosen; other reasonable bases give anywhere from about 40 to 70 times, which is itself a sign of how thin the margin is.

The historical record: confirmed and sourced. Our claim that no company with sales above $50 billion grew faster than 35 percent a year over five years, from 1950 to 2015, comes from The Base Rate Book (Exhibit 4, page 25). Of 1,810 five-year periods for companies of that size, the fastest fell between 30 and 35 percent, and only one did; the median was 1.5 percent a year after inflation. Over three years, two periods reached 40 to 45 percent, which is why our comparison is over five.

The head start: confirmed, and two sentences cut. If Anthropic triples its revenue next year, the spider then needs 27 percent a year for four years and the shark 54 percent. Claude had recomputed the shark at 53; Shiva showed that Table 3 is built on unrounded multiples (about 10.6 times for Netflix, not 11), on which 54 is correct. We also cut two sentences about the growth needed if investors accept no return at all, because they switched from one test to another without saying so and confused a careful reader.

Data, as of late October 2026: StockAnalysis.com for Palantir’s and CoreWeave’s market values and ratios; CoreWeave’s Form 10-Q for the second quarter of 2026 and its August guidance; Michael Mauboussin, Dan Callahan and Darius Majd, The Base Rate Book (Credit Suisse, 2016). Calculations were rerun in Python.

This piece is co-authored with Prof. Dr. Kerstin Lopatta, an accounting professor at the University of Hamburg.

Quinten Metsys, The Moneylender and His Wife, 1514. Musée du Louvre, Paris. Public domain, via Wikimedia Commons.

Last week we published a technical note on the requirement in the European Sustainability Reporting Standards (ESRS) to disclose the “anticipated financial effects” of climate-related risks and opportunities. It was written for accountants and others who work on these questions, and it was deliberately dry. One distinction at the center of it deserves a wider audience, because it bears on almost every argument about forward-looking climate information, whoever is making the argument.

The distinction is between two ways the future can enter a company’s accounts. In the first, the company owes something, or owns something, today, and the future matters because it determines how much that thing is worth. In the second, the company does not yet owe or own anything, and the future matters because of what may happen to its business. Financial accounting has a long and well-developed practice for the first. For the second it has a deliberate answer, which is that such information belongs somewhere other than the financial statements. Much of the current debate about climate reporting turns on which of the two a particular request falls under.

A Hanseatic merchant in London among his letters, scales and accounts, under his motto “No pleasure without sorrow.” Hans Holbein the Younger, The Merchant Georg Gisze, 1532. Gemäldegalerie, Staatliche Museen zu Berlin. Public domain, via Wikimedia Commons.

How the Future Gets into the Accounts

It is a common misconception that financial statements are purely historical. They are full of estimates about the future. A company with a defined benefit pension plan measures what it will owe retirees decades from now. An oil company measures the cost of dismantling a platform at the end of its life, perhaps thirty years away. A life insurer measures claims on policies that will run for the rest of its customers’ lives. A bank sets aside allowances for loans that may default years from now, using forecasts of the economy.

None of this is controversial in principle, because in each case there is an anchor. The pension promise has already been made. The platform has already been built, and the obligation to remove it has already arisen. The insurance policy has already been written. The loan has already been made. The IASB’s Conceptual Framework states the principle in paragraph 3.6: forward-looking information is included in the financial statements only if it relates to the entity’s assets or liabilities, including unrecognized ones, or equity, that existed at the end of the period or during it, or to income and expenses for the period, and if it is useful to users of the statements. The horizon of the estimate can be very long, but it follows from something that exists now.

The anchor does not make these estimates easy, or free of judgment. A sixty-year pension obligation depends on the forecaster as much as any projection does. What the anchor provides is something else, and it is worth being precise about it. First, it provides a unit of account, an item whose existence can be verified at the reporting date: a promise made, a platform built, or a policy written. Second, it provides limits: the carrying amount of an asset bounds how much of it can be written off. Third, it allows standardized methods, such as the requirement in IAS 19 to discount pension obligations at the yield on high-quality corporate bonds, which make estimates by different companies more comparable. Fourth, it leads to realization: the pension is eventually paid, the platform removed, the claim settled, and the estimate can be tested against what happened. These four properties, not the length of the horizon, are what distinguish accounting estimates from forecasts in general.

The same principle governs how accounting treats broader trends. When a company tests a power station for impairment, it estimates the cash flows the plant will generate, and those estimates depend on assumptions about electricity prices, fuel costs and, increasingly, carbon prices. A macro trend therefore enters the accounts directly. But it enters as an assumption used to value something the company already has. If the expected carbon price rises, the plant may be worth less, and the company records an impairment. The trend changes the value of an existing asset. It does not, by itself, create anything new to put on the balance sheet.

The effect is also asymmetric. Under the cost model, impairment works downward. An asset is written down when its recoverable amount falls below its carrying amount, and a later reversal cannot take it above what its carrying amount would otherwise have been. A trend that raises value, such as the falling cost of renewable technology for a company positioned to use it, does not, for assets measured at cost, show up in the accounts at all until it produces revenue. Anticipated financial effects, as the sustainability standards define them, cover opportunities as well as risks, and this prudence asymmetry is one reason investors ask for such information outside the financial statements.

Accounting also disciplines how such projections are made. When management estimates the value in use of an existing business, IAS 36 limits the projections based on detailed budgets and forecasts to five years unless a longer period can be justified, after which cash flows are extrapolated at a steady or declining growth rate. The projections must also exclude the benefits of restructurings to which the company is not yet committed and of investments that would improve the asset’s performance. Through the terminal value, value in use still reaches far into the future, and fair value less costs of disposal, the alternative measure of recoverable amount, has no such limit. The discipline lies less in the length of the horizon than in rules about what may be assumed.

So the principle can be stated simply. A trend can inform the measurement of something a company already has or already owes. It cannot, on its own, create something to measure.

Two officials record what the law says is owed. Marinus van Reymerswaele, Two Tax Gatherers, c. 1540. Public domain, via Wikimedia Commons.

Where Obligations Come From

If obligations are the anchor, the next question is where obligations come from. Here accounting offers something close to a spectrum, and it is worth walking along it, because climate-related requests fall at different points on it.

Contracts. The clearest obligations are the ones a company has signed. A long-term agreement to buy renewable power, a commitment to deliver low-carbon products to a customer, or a loan covenant tied to emissions creates rights and obligations that exist today. Existence is not the same as recognition, however. Many such contracts are executory, meaning that neither party has yet performed, and accounting generally does not recognize them unless they become onerous or fall within the rules for derivatives or leases. A power purchase agreement can be firmly anchored without appearing on the balance sheet. It is still a unit of account, and accounting knows how to describe it.

Enacted law. Legislation creates obligations without the company’s consent, and accounting has precise rules for when it does so. Under IAS 12, income taxes are measured using rates that have been enacted or substantively enacted by the reporting date. In some jurisdictions a government announcement can have the substantive effect of enactment, but a rate change that has merely been announced is generally not used. For levies imposed by governments, IFRIC 21 provides that the obligation arises when the company carries out the activity that the legislation identifies as triggering payment. However, the IASB has proposed withdrawing the interpretation as part of targeted improvements to IAS 37, so this treatment may change. A company covered by the EU Emissions Trading System incurs an obligation to surrender allowances as it emits. There is no IFRS Standard dealing specifically with emission allowances, since IFRIC 3 was withdrawn in 2005, and practice varies. But under the approaches in common use, that obligation is real, measurable and auditable, because the law that creates it is in force.

Constructive obligations. A company can also create an obligation through its own conduct. Under IAS 37, a constructive obligation arises when a company has indicated to others, through an established pattern of practice, published policies or a sufficiently specific current statement, that it will accept certain responsibilities, and has thereby created a valid expectation that it will do so. In April 2024 the IFRS Interpretations Committee considered how this applies to a public climate commitment. Its conclusions illustrate the anchor principle precisely. A company’s public commitment to reduce its emissions and offset what remains could create a constructive obligation. But making the commitment does not, by itself, require a provision, because no past event has yet occurred. The costs of reducing future emissions are costs of operating in the future, recognized when they are incurred. Once the company emits in a period it has committed to offset, however, it has a present obligation to buy carbon credits for those emissions, and a provision is recognized. The same commitment produces no liability for future operations and a measurable liability for past emissions. The Committee also noted that the actions a company plans in order to fulfill such a commitment may affect the measurement of other assets and liabilities, and the related disclosures, whether or not a provision is recognized. A commitment that creates no liability can still change the accounts through the assets it affects.

Possible obligations. Between what a company clearly owes and what it does not lies a category accounting handles by disclosure. IAS 37 defines a contingent liability as a possible obligation arising from past events whose existence will be confirmed only by uncertain future events, or a present obligation that is not recognized because an outflow is not probable or cannot be measured reliably. Such liabilities are disclosed unless the possibility of an outflow is remote. Climate litigation often falls here. Claims that a company should pay for harm caused by its past emissions rest on past events, but whether they create an obligation depends on law that is still being made. The anchor exists; the obligation is uncertain.

Announced policy. Governments regularly announce intentions that have not yet become law: a future carbon tax, a phase-out date for a technology, a planned tightening of standards. These create no obligation. They can, and often should, inform the assumptions used to measure existing assets, as in the power station example. For income taxes, accounting generally waits for substantive enactment before using an announced rate.

International agreements. The Paris Agreement and similar instruments bind states, not companies, and even for states their central commitments are nationally determined. They reach a company’s accounts only when a government translates them into domestic law that applies to the company, as the European Union has done through instruments such as the Emissions Trading System. They can shape expectations about future policy, and those expectations can inform estimates. They do not create an obligation for any company.

Macro trends. At the far end of the spectrum are developments that no one has committed to: the physical effects of a warming climate, the falling cost of renewable technologies, shifts in consumer preferences, changes in the cost of capital for carbon-intensive activities. These may be the most consequential forces of all for a company’s long-term prospects. In accounting terms, they never create obligations. They enter the financial statements only as assumptions in the measurement of things that already exist.

Accounting recognizes obligations from the first three points on this spectrum, when the conditions for recognition are met, and discloses possible obligations from the fourth. It uses the last three to inform estimates, and for tax rates it generally waits for enactment before using even announced policy.

A geographer pauses over his chart, dividers in hand, and looks toward the window. Johannes Vermeer, The Geographer, 1669. Städel Museum, Frankfurt. Public domain, via Wikimedia Commons.

Three Kinds of Forward-Looking Climate Information

The spectrum suggests a way of sorting the forward-looking climate information that investors, regulators and others ask companies to provide. The dividing line is whether the information has a unit of account: whether it is about something that exists at the reporting date. On that test, and with one further distinction, it falls into three kinds.

The first is information about obligations and rights that exist today, whether or not they are recognized: the obligation to surrender emission allowances, a decommissioning liability, a provision to buy offsets for past emissions, a long-term power purchase agreement, a contingent liability arising from climate litigation. Information of this kind can be measured or described, compared and audited in the same way as other accounting information, and most of it is already in the financial statements or the notes.

The second is information about how trends affect the measurement of existing items: the carbon price assumed in an impairment test, the climate scenarios used in a bank’s expected credit loss model, the sensitivity of a recoverable amount to a change in a key assumption, the reduced value of existing installations along a particular climate pathway. Value in use is itself a forecast of the future operations of an existing unit, so information of this kind can reach far into the future; what places it here is that it measures something that exists. Exposure metrics belong here too, such as the carrying amount of assets in areas exposed to physical hazards. The carrying amount is anchored, but classifying an asset as exposed rests on hazard models and definitions that are not yet standardized, and making such metrics comparable will take real work. Much of this information is already part of financial reporting. Accounting standards require disclosure of key assumptions and of the sources of estimation uncertainty that could materially change carrying amounts within the next year, and the IASB has published illustrative examples of how climate-related matters should be reflected in those disclosures. The practical question here is not whether such information should be provided, but whether it is being provided well.

The third is information with no unit of account at all: the revenue a company expects to gain or lose in markets it does not yet serve, the costs of operations it has not yet committed to, the returns on investments it has not yet made, the overall change in its financial position as markets shift. This information can be highly relevant to an investor’s assessment of a company’s prospects. Like any estimate, it depends on the forecaster. But it lacks all four of the properties the anchor provides. Nothing in it can be verified to exist at the reporting date; no carrying amount bounds it; no standardized method governs it; and later outcomes rarely isolate the effect the estimate was about, so it can seldom be tested against what happened. Financial accounting sends information of this kind to management commentary, and in practice the most useful form for it is conditional: what would happen under stated scenarios, presented as such.

A woman holds a balance at rest, its pans empty, beneath a painting of the Last Judgment. The painting was long known as Woman Weighing Pearls, until close examination showed the pans hold nothing. Johannes Vermeer, Woman Holding a Balance, c. 1664. National Gallery of Art, Washington. Public domain, via Wikimedia Commons.

What the Distinction Implies

The sorting does not settle what companies should be required to disclose. That is a policy question, and reasonable people answer it differently. It does clarify what different requests are asking for and what discipline can be applied to the answers.

Requests for more forward-looking climate information often concern the third kind, because that is where long-term climate effects mostly lie. Such requests are not unreasonable; investors legitimately want this information, and analysts already make such projections themselves. The technical point is that information of the third kind cannot be given the properties of accounting information simply by being presented in monetary amounts or connected to line items in the financial statements. The anchor does not make an estimate reliable. It supplies a verifiable unit of account, a limit, standardized methods and eventual realization, and trends supply none of these. If information of the third kind is to be useful and comparable, the conventions that govern it have to be developed on their own terms: explicit scenarios, stated assumptions, ranges rather than point estimates, separation of what management expects from what would happen under a specified pathway, and assurance that addresses the process by which the numbers were produced rather than whether the forecasts will come true.

There is a precedent for that last element. German law has long required a forecast report in the management report: under the German Accounting Standard DRS 20, companies forecast their most significant performance indicators for at least the following year, and under § 317(2) of the Commercial Code the auditor examines whether the opportunities and risks of future development are presented appropriately. The auditor assesses the plausibility of the forecast and the process behind it, not whether it comes true.

The opposite position, that companies should provide no forward-looking climate information at all, does not fit the accounting either. Financial statements already contain a great deal of forward-looking climate-relevant estimation, because climate trends affect the value of existing assets and liabilities now. A company that holds long-lived assets exposed to carbon prices or physical hazards cannot measure those assets faithfully without making assumptions about the future, and the standards require it to disclose the key ones.

Financial accounting sits between these positions, and it has occupied that ground for a long time. It admits the future freely when the future determines the value of something that exists today, however distant the horizon. It treats the future of a company’s own operations with caution, disciplining projections and excluding what has not been committed. And it does not treat a trend as an obligation, however strong the trend, until something converts it into one: a contract, a law, or the company’s own commitment.

The sample masters of the Amsterdam drapers’ guild, who inspected cloth and certified that it met the guild’s standards, look up as if a question has just been put to them. Rembrandt van Rijn, The Syndics of the Drapers’ Guild (De Staalmeesters), 1662. Rijksmuseum, Amsterdam. Public domain, via Wikimedia Commons.

A Question for Standard-Setters

For those who write sustainability reporting standards, the distinction raises a practical question. Both the ISSB’s IFRS S2 and the revised ESRS already distinguish current financial effects from anticipated ones, which corresponds broadly to the line between our first two categories and the third. And in the European Union the sustainability statement is already part of the management report, not the financial statements. The question is therefore not where the information goes. It is the form in which it is requested.

Three requirements of the revised ESRS show what we mean. Paragraph 27 of ESRS 2 asks a company to disclose how it expects its financial position, financial performance and cash flows to change over the short, medium and long term, given its strategy: information of the third kind, requested in the vocabulary of the financial statements. IFRS S2 asks for similar information. Paragraph 40(d) of ESRS E1 asks for estimated potential liabilities related to climate transition that do not meet the recognition criteria at the reporting date but may have to be recognized in future periods, which are amounts that accounting has decided not to recognize. And paragraph 40(a) asks for a range of potential stranded assets under a scenario aligned with limiting warming to 1.5°C. That is measurement of existing assets, our second category, but under a scenario rather than management’s expectations, and it is presented alongside carrying amounts as though it were an expectation. The standards soften these requirements in two ways. First, reliefs modeled on those in IFRS S1 allow a company to omit quantitative information where the effects cannot be separately identified, where measurement uncertainty is too high, or where it lacks the necessary capabilities. Second, phase-ins defer most quantitative information on anticipated effects to 2030 for the largest companies. But reliefs and phase-ins postpone the question of form rather than answer it.

The alternative is to develop, deliberately and explicitly, a separate set of conventions for conditional, trend-based information, designed for what it is. Some of those conventions already exist in supervisory scenario analysis, in the sensitivity disclosures of financial reporting and in forecast reporting of the German kind. How far they can be adapted for general purpose reporting to investors is, we think, one of the more interesting technical questions in the field, and one that does not depend on where anyone stands on climate policy.

About the Authors

Robert G. Eccles is Visiting Professor of Management Studies at the Saïd Business School, University of Oxford.

Kerstin Lopatta is Professor of Financial Accounting, Auditing and Sustainability at the University of Hamburg.

The views expressed are the authors’ own.

Laurence Sterne by Joshua Reynolds, 1760. © National Portrait Gallery, London. Pubic domain via Wikimedia Commons.

Laurence Sterne’s The Life and Opinions of Tristram Shandy, Gentleman, published in nine volumes between 1759 and 1767, is the story of a man who sets out to tell the story of his life and cannot manage to get himself born. His father’s theories, his uncle’s hobby-horse, and a long series of digressions intervene. The hero does not arrive until the third volume. When he does, the arrival goes badly. The forceps of Dr. Slop, the man-midwife, flatten his nose, and the name his father chose with great care, Trismegistus, is garbled on the way to the christening into Tristram, the one name his father thought worst of all. No one in the household means him harm. Everyone is trying to help.

I have been thinking about Tristram because I have been trying to understand ESRS-40a, the sustainability reporting standard the European Union is preparing for large non-EU companies that do business in Europe. EFRAG’s public consultation on it closes on October 31—three days from today. The questions I set out to answer seemed simple. What is this standard? Who is making it? What will it require, and of whom? And once it exists, how firmly will it hold? Each answer turned out to depend on something that happened earlier, in a different piece of legislation, for a different reason. So, in the Shandean manner, this piece proceeds by digression. Readers who know their way around EU law can skip the digressions. Everyone else may find them the most useful part.

A Digression on Directives, Regulations, Articles and Recitals

Digressions, incontestably, are the sunshine;—they are the life, the soul of reading!

— Laurence Sterne, Tristram Shandy

For American readers, a rough analogy helps. Think of the European Parliament, whose 720 members are elected directly by voters in the 27 member states, and the Council of the European Union, where ministers from the 27 national governments sit, as the two houses of Congress. Because both must agree before a law is passed, they are called the “co-legislators.” Think of the European Commission, the EU’s executive, as something like a federal agency, with one important difference. Only the Commission can propose legislation; the co-legislators amend it and adopt or reject it.

What Congress does with statutes, the European Union does with directives and regulations. A regulation applies directly in every member state. A directive sets out what must be achieved and leaves each member state to write it into its own law, a process called transposition. The rules on corporate reporting live in a directive, the Accounting Directive of 2013, which has since been amended many times. In European shorthand, the directive is “Level 1.”

Directives can delegate rulemaking power to the Commission, much as a statute delegates it to the SEC. The rules the Commission adopts under that power are called delegated acts, and they are “Level 2.” A delegated act may supplement the directive but may not override it, and the Parliament and the Council each have two months to veto it, much as Congress can under the Congressional Review Act. The European Sustainability Reporting Standards, or ESRS, are delegated acts. So will ESRS-40a be.

Directives and delegated acts are divided into numbered articles, which are the law, and preceded by numbered recitals, which are not. Recitals explain what the legislators were trying to do. Courts read them when the articles are unclear, and they are often the most candid part of the text.

Finally, because a directive amended many times becomes very hard to read, the EU’s Publications Office produces a “consolidated version” that stitches the amendments into the original. The current one, dated March 18, 2026, opens with a warning: “This text is meant purely as a documentation tool and has no legal effect.” The warning matters more than it appears to, as we shall see.

Title page of the first volume of Tristram Shandy, printed by Ann Ward, York, 1760. The epigraph, from Epictetus, reads: “Men are disturbed not by things, but by their opinions about things.” Public domain, via Wikimedia Commons.

Conception: The CSRD and Article 40a

I wish either my father or my mother, or indeed both of them, as they were in duty both equally bound to it, had minded what they were about when they begot me.

— Laurence Sterne, Tristram Shandy

The Corporate Sustainability Reporting Directive, or CSRD, was adopted in December 2022 as a set of amendments to the Accounting Directive. It was conceived in a spirit of great ambition. It would cover, by the Commission’s estimate, some 50,000 companies, including all large companies and listed small and medium-sized ones, public or private. It would require “double materiality,” meaning reporting both on how sustainability matters affect a company and on how the company affects people and the environment. And the first set of standards adopted under it, in 2023, contained around 1,200 datapoints, already cut from some 2,500 in EFRAG’s 2022 drafts. Many in the NGO community saw it as the moment corporate reporting would finally be enlisted in changing the world.

The CSRD also reached beyond Europe. It inserted a new Article 40a into the Accounting Directive for groups headquartered outside the EU with significant business inside it. Recital 20 of the CSRD explained the purpose: to ensure that such groups “are accountable for their impacts on people and the environment and that there is a level playing field for companies operating in the internal market.” As enacted, Article 40a applied to non-EU groups with more than €150 million of EU turnover and an EU subsidiary or branch of a certain size.

The mechanism was ingenious, and it is the first thing to understand about ESRS-40a. The EU cannot impose obligations on a company in Ohio. It can impose them on that company’s subsidiary in Ireland. So Article 40a requires the EU subsidiary (or, if there is none, the branch) to “publish and make accessible a sustainability report... at the group level of that ultimate third-country parent undertaking.” The legal duty falls on the European child; the report covers the whole family. The content is confined to impacts, with no reporting on the financial risks and opportunities that sustainability matters create for the group, but it covers every sustainability topic, from climate to workforce conditions to governance, and includes the group’s transition plan and targets.

Article 40a sets out who must report and, in broad terms, what the report must cover. It leaves the detail to a companion provision, Article 40b, which told the Commission to adopt a delegated act, by June 30, 2024, setting out standards that specify exactly what information the report must contain. That standard is what EFRAG is now drafting. It is called ESRS-40a after Article 40a, which creates the obligation, although it will be adopted under Article 40b, which creates the power to write it. Keep both numbers in mind; a third, Article 49(3b), is about to join them.

A Digression on EFRAG and DG FISMA

The Commission does not write sustainability reporting standards itself. It relies on EFRAG, the European Financial Reporting Advisory Group, which the CSRD, in its recitals, describes as “a non-profit association established under Belgian law.” EFRAG began life advising the Commission on international accounting standards and added a sustainability reporting pillar in 2022, governed by a Sustainability Reporting Board and supported by a technical expert group and a secretariat. EFRAG’s Administrative Board oversees its governance and due process. Its sustainability reporting work is co-financed by the EU budget, which covers up to 90 percent of the eligible costs, and by its member organizations.

On the Commission’s side, the responsible department is the Directorate-General for Financial Stability, Financial Services and Capital Markets Union, known as DG FISMA, which reports to the Commissioner for Financial Services and the Savings and Investments Union, Maria Luís Albuquerque. In practice the relationship works like this. DG FISMA asks EFRAG for technical advice, EFRAG develops it through its own consultations and board deliberations, and DG FISMA turns that advice, with whatever changes the Commission decides to make, into a delegated act that the Commission adopts. EFRAG advises; the Commission decides.

The marbled page from Volume III of Tristram Shandy, 1761, which Sterne called “the motly emblem of my work.” Public domain, via Wikimedia Commons.

A Digression on Article 49(3b) and a Missing Sibling

How much weight must the Commission give EFRAG’s advice? The CSRD answered that question in a provision it inserted as Article 49(3b) of the Accounting Directive. As enacted in 2022, it read:

“When adopting delegated acts pursuant to Articles 29b and 29c, the Commission shall take into consideration technical advice from EFRAG, provided that: (a) such advice has been developed with proper due process, public oversight and transparency, with the expertise and balanced participation of relevant stakeholders, and with sufficient public funding to ensure its independence…”

Further conditions follow, on cost-benefit analysis and on who may take part in EFRAG’s work, along with a requirement that the Commission seek the opinions of the European supervisory authorities and consult several other EU bodies.

Article 29b is the Commission’s power to adopt the main ESRS. Article 29c was its power to adopt a separate, lighter set of standards for listed small and medium-sized companies, which EFRAG duly drafted and published for consultation in January 2024. Article 40b, the power to adopt standards for non-EU groups, is not mentioned. The co-legislators wrote EFRAG’s advisory role, and the conditions attached to it, into the law for the main standards and for their small-company sibling. They never wrote them in for the non-EU standards. Article 40b itself says only that the Commission shall adopt “a delegated act... to provide for sustainability reporting standards for third-country undertakings that specify the information that is to be included in the sustainability reports referred to in Article 40a.” It does not mention EFRAG at all.

EFRAG nonetheless went on to draft standards under all three powers, because the Commission asked it to. The listed-company standards under Article 29c never made it into law. In 2026, the Omnibus I Directive, the law that carried out the EU’s simplification of the CSRD (described below), deleted Article 29c altogether. Since that power no longer existed, Omnibus I also removed it from Article 49(3b), whose opening now reads “When adopting delegated acts pursuant to Article 29b.” So the sentence that once covered two sets of standards now covers one, and it never covered the non-EU standards at all. A reader of today’s consolidated text sees “Article 29b” standing alone, marked as amended, with no trace of the sibling that once stood beside it. It would be natural to assume that a provision guaranteeing due process for EFRAG’s advice governs every standard EFRAG writes. It governs one.

The consequence is that, as a matter of law, EFRAG’s advice on ESRS-40a has no special standing. The Commission may adopt it unchanged, change it as it likes, or, in principle, write the standard without it. Its legal constraints are the general ones. It must stay within what Articles 40a and 40b permit, consult experts designated by the member states, give reasons, and survive the two-month window in which the Parliament or the Council may object. EFRAG is drafting ESRS-40a by invitation, not by statute.

Dr. Slop, the man-midwife, slumped at left, with Walter Shandy, Uncle Toby and Corporal Trim, who reads a sermon. Engraving after William Hogarth’s frontispiece to Tristram Shandy, 1760. Wellcome Collection, CC BY 4.0.

The Forceps

…thy new-invented forceps,—thy crotchet,—thy squirt, and all thy instruments of salvation and deliverance…

— Laurence Sterne, Tristram Shandy

The ESRS-40a that EFRAG is now drafting is not the standard the CSRD’s authors imagined, because the law around it changed shape several times before it could be born.

The first change was simply time. In 2024, Directive (EU) 2024/1306 pushed the Article 40b deadline back two years, to June 30, 2026, to give EFRAG and companies more room. The second was politics. By 2025 the European mood had shifted from ambition to competitiveness. The Commission proposed its “Omnibus” simplification package in February 2025, the co-legislators passed a “stop-the-clock” directive that April to delay reporting for many companies, and the Omnibus I Directive, Directive (EU) 2026/470, was adopted in February 2026. It confined the CSRD to EU companies with more than 1,000 employees and €450 million of turnover, removing roughly 90 percent of the companies previously covered. For non-EU groups it tripled the EU turnover threshold to €450 million in each of the last two years, raised the size test for the EU subsidiary or branch to €200 million of turnover, and let groups headed by certain financial holding companies opt out entirely. EFRAG estimates that about 1,200 non-EU groups remain in scope, against some 10,000 under the original rules. In July 2026 the Commission adopted the revised ESRS for EU companies, with about 60 percent fewer mandatory datapoints.

The third pressure came from across the Atlantic. In the Joint Statement on the EU–U.S. trade framework of August 2025, the European Union committed “to undertake efforts to ensure that the Corporate Sustainability Due Diligence Directive (CSDDD) and the Corporate Sustainability Reporting Directive (CSRD) do not pose undue restrictions on transatlantic trade.” The United States has since made clear that it does not think the Omnibus went far enough. I will return to that in a separate piece.

Like Dr. Slop’s forceps, each intervention was well meant, and each left its mark on the shape of what was eventually delivered.

I am this month one whole year older than I was this time twelve-month; and having got, as you perceive, almost into the middle of my third volume…—and no farther than to my first day’s life…

— Laurence Sterne, Tristram Shandy

The Child Takes Shape

The exposure draft for ESRS-40a follows the architecture of the revised ESRS but, as Article 40a requires, deals only with impacts. It offers a company three choices: report on all its material impacts globally; apply the full ESRS voluntarily; or use a “mixed approach,” under which climate impacts are always reported globally but, for other topics, reporting may be confined to impacts connected with the EU, where these can be meaningfully separated from the rest.

The mixed approach has an unusual parentage. EFRAG has said that it was included at the Commission’s request, and its Sustainability Reporting Board approved the draft with reservations about the level playing field, comparability, and the feasibility of assurance, and, according to Ropes & Gray, about “whether there is a legal basis for the mixed approach.” That doubt is easy to understand once the two articles are set side by side. Article 40a(1) requires a report “at the group level” of the non-EU parent. Article 40b empowers the Commission to “specify the information that is to be included” in that report. The Directive itself decides whose activities the report covers: the whole group’s. The Commission’s power is only to decide what information goes into it. Allowing a group to confine most of its reporting to its EU-related impacts looks less like deciding what information to include than like shrinking the group the report covers, which the Directive has already defined.

The consultation through which EFRAG is gathering views is also distinctive. It runs for 100 days rather than the 120-day minimum in EFRAG’s own due process procedures, under an expedited process the Administrative Board approved in April without recording its reasons. Responses are accepted only through an online questionnaire, with a 300-word limit per answer and no provision for comment letters, although EFRAG accepted letters when it consulted on the revised ESRS a year earlier. Publication of the responses depends on each respondent’s consent, and no date has been set.

By comparison, the International Accounting Standards Board and the International Sustainability Standards Board normally allow at least 120 days, accept comment letters and post them as they arrive. This week the ISSB opened a consultation on a nature-related practice statement, which companies need not even apply, with 120 days and two channels for comment. ESRS-40a, which will be binding EU law, has 100 days and one. As it happens, EFRAG’s consultation on its own agenda for 2027 to 2029, opened on October 16, accepts comment letters, publishes them, and promises a feedback statement explaining how the responses shaped its priorities. ESRS-40a gets none of these.

Time wastes too fast: every letter I trace tells me with what rapidity Life follows my pen…

— Laurence Sterne, Tristram Shandy

Each of these features is within EFRAG’s procedures. But because Article 49(3b) does not apply, EFRAG’s consultation is the only substantial due process ESRS-40a will receive before it reaches the Commission, and it is the one that has been compressed. What happens next is harder still to see. Between October 31 and January, EFRAG will analyze the responses and finalize its advice. Its board meets in public and posts its papers, but what those papers will reveal about the responses, and whether and when the responses themselves will be published, no one outside EFRAG knows. At the IASB and the ISSB, comment letters are posted as they arrive, so anyone can read them while a consultation is still open. For ESRS-40a, if the responses appear at all, they will appear all at once, possibly after the advice they informed has already been delivered. The Commission’s part in that period is informal and unrecorded. After January, DG FISMA will draft the delegated act. EFRAG has said that the Commission “will subsequently launch its own public consultation before adopting the delegated act” (EFRAG Update, July 2026), as the Commission’s own Better Regulation practice provides, but nothing in the Directive obliges it to, or to explain how EFRAG’s advice, or the responses behind it, shaped the result.

The christening of Tristram: Walter Shandy arrives in his nightgown, too late to stop the curate from giving his son the wrong name. Engraving after William Hogarth’s frontispiece to Volume IV of Tristram Shandy, 1761. The Metropolitan Museum of Art, public domain.

A Digression on the Name

How many Caesars and Pompeys, he would say, by mere inspiration of the names, have been rendered worthy of them?

— Laurence Sterne, Tristram Shandy

Tristram’s father believed that names shape destinies, and lost his son’s to a maid who could not carry two long words down a corridor. ESRS-40a’s name, as noted earlier, comes from Article 40a, although the standard will be adopted under Article 40b. That is perfectly sensible, and it is also a small emblem of the whole story: a standard whose identity is borrowed from one article (40a), whose legal footing lies in another (40b), and whose governance was settled, or rather left unsettled, in a third (49(3b)).

Sterne’s diagram of the plot lines of the first four volumes of Tristram Shandy, from Volume VI, chapter 40, 1762, “invented” and “engraved” by Tristram himself. Having charted his digressions, he resolves to do better: “by the help of a vegetable diet, with a few of the cold seeds, I make no doubt but I shall be able to go on with my uncle Toby’s story, and my own, in a tolerable straight line.” Public domain, via Wikimedia Commons.

What Will It Be?

Suppose ESRS-40a is born on schedule. How firmly will it hold? Here the text of Article 40a holds a surprise, and the co-legislators themselves describe it.

The EU subsidiary must ask its non-EU parent for the information it needs. If “not all the required information is provided,” it must publish “all information in its possession, obtained or acquired,” together with “a statement indicating that the third-country undertaking did not make the necessary information available.” The report must carry an assurance opinion, but if the parent does not provide one, the subsidiary again issues a statement saying so. Its directors are responsible for the report “to the best of their knowledge and ability.” The reports themselves are filed nationally, in each member state’s business register, not with any EU authority. Member states “may,” not must, tell the Commission which subsidiaries published a report and which used the non-cooperation statement, and the Commission publishes on its website a list of the non-EU groups that do report. Enforcement is therefore a national matter, and member states vary considerably in how promptly and how vigorously they apply EU law. Six of them, including Germany and the Netherlands, have still not fully transposed the original CSRD into national law, more than two years after the deadline (Gibson Dunn). Recital 20 of the CSRD framed all of this as what follows “despite the best efforts of the subsidiary.” Recital 26 of Omnibus I is blunter: the subsidiary “is only required to publish and make available the sustainability report provided by the third-country undertaking, whereas undertakings subject to Articles 19a and 29a are required to report on their own behalf.”

In other words, the legal duty falls on a conduit. The non-EU parent is never bound, and a parent that declines to cooperate exposes its subsidiary not to a penalty but to a public statement. That is not nothing. The statement is filed in a public register, the Commission’s list is public, and a subsidiary that already holds much of the group’s information, through shared systems or the parent’s own sustainability reporting, may find “in its possession, obtained or acquired” a more demanding phrase than it sounds. But ESRS-40a’s force will be largely reputational.

The firmer obligations lie elsewhere. A non-EU group’s EU subsidiaries that are large enough in their own right, with more than 1,000 employees and €450 million of turnover, must report on their own behalf under the full ESRS, double materiality included, and there is no non-cooperation exit for them. A transitional provision, Article 48i, lets such subsidiaries report together in one EU-level report until January 6, 2030. After that, each reports separately unless they sit under a single EU holding company or the parent reports for the whole group under the full ESRS or an equivalent standard. A report under ESRS-40a does not qualify.

The post-2030 rule deserves a moment’s reflection. EU subsidiaries already publish their own financial statements, so separate reporting as such is not new. But financial statements work at the level of the legal entity because transactions are booked entity by entity; there is a ledger, and its boundary is clear. Sustainability impacts do not respect those boundaries. Supply chains are managed centrally, and procurement, product design and climate strategy are decided at group level. Financial reporting has long recognized that legal entities can be a poor guide to how a business is actually run: segment reporting under IFRS 8 and its US equivalent follows management’s organization, not the corporate chart. The post-2030 rules do the reverse, asking each large subsidiary to report as if it were self-contained on impacts that are largely decided and managed elsewhere in the group. Faced with a separate report for each, many groups may find it simpler to report globally under the full ESRS, which may prove a more effective form of extraterritorial reach than Article 40a itself.

Uncle Toby, under his map of the fortifications of Dunkirk, peers into the Widow Wadman’s eye for a speck that isn’t there. She is laying siege to him; his hobby-horse is the art of siege itself. Charles Robert Leslie, My Uncle Toby and the Widow Wadman, 1831. Victoria and Albert Museum, London. Public domain, via Wikimedia Commons.

Who Should Worry?

…so long as a man rides his Hobby-Horse peaceably and quietly along the King’s highway, and neither compels you or me to get up behind him,—pray, Sir, what have either you or I to do with it?

— Laurence Sterne, Tristram Shandy

The answer, curiously, is that both sides of the long argument over the CSRD may have less at stake in ESRS-40a than they think.

Non-EU companies, many of them American, have treated Article 40a as a sweeping assertion of extraterritorial authority. On paper it reaches the global group. In practice the legal duty rests on a subsidiary that must try, and a parent that declines faces disclosure rather than compulsion. The law binds the subsidiary, but it cannot bind the parent, so whether the information flows is in practice the parent’s decision. The mixed approach reduces the burden mainly for groups whose parents decide to supply it. The obligations with real force are the ones on their large EU subsidiaries, under the main ESRS, and those are not what ESRS-40a is about.

NGOs, who saw in Article 40a a way to hold the world’s largest companies “accountable for their impacts on people and the environment,” have already watched the threshold triple, the population of covered groups fall by nearly nine-tenths and a holding-company exemption appear. They are now looking at a standard developed without the statutory due process that governs the main ESRS, with a mixed approach that would let groups confine most of their reporting to Europe, and with a legal mechanism that allows a parent to decline. Many of them are already dismayed by Omnibus I. ESRS-40a may disappoint them further, and most have not yet noticed.

What the standard will be, therefore, depends less on its text than on who chooses to test it, and both sides have reasons to.

Laurence Sterne, a version of Joshua Reynolds’s 1760 portrait attributed to William Etty, with the papers of a manuscript at his elbow. Art UK, public domain, via Wikimedia Commons.

An Unfinished Book

L—d! said my mother, what is all this story about?——A Cock and a Bull, said Yorick——And one of the best of its kind, I ever heard.

— Laurence Sterne, Tristram Shandy

Sterne published the ninth and last volume of Tristram Shandy in 1767 and died the following year. The book ends without its hero having told us much of his life at all; the opinions crowd out the life. ESRS-40a is in a similar condition. It was conceived in ambition, delayed by necessity, reshaped by politics, governed by a provision that turns out not to apply to it, and endowed with a legal mechanism its own legislators describe as publication of what someone else provides. It is due to be delivered in January, to be adopted by an institution that may change it as it likes, and to apply to companies that may, in law, decline to help.

None of this is anyone’s villainy. It is what happens when a standard is shaped by successive acts of legislation, each written for its own reasons, by people who were not always looking at what the others had done. The result is a standard whose purpose, content, process and force are all, for now, open questions. The consultation closes on October 31. What EFRAG makes of the responses, what the Commission makes of EFRAG’s advice, and what anyone will be able to see of either, will tell us a good deal more about what ESRS-40a is. Whether the European Union will still want it to be anything at all is a question I take up in the next piece, on the CSRD and the trade negotiations with the United States.

This piece is co-authored with Shivaram Rajgopal and available on The Googly.

In September we published our analysis of the risks we think matter most to Anthropic’s value, and our predictions of how its IPO prospectus will describe them. We then asked readers to do the same exercise in a simpler form. The survey listed 17 risks, asked respondents to rate how likely each was to materially reduce the value of Anthropic’s shares over the next three years; we also invited them to tell us what we had missed. It ran from October 1 to October 20, 2026, and 51 people answered. This piece reports what they said.

We do not want to make too much of it. Fifty-one is a small sample, the respondents chose themselves, and most are not specialists in Anthropic or in AI companies. Then again, few people are, and the market itself is only beginning to work out how to assess the risks of a company like this one. The survey is a snapshot taken at a particular moment, by a method quite different from the one behind our September piece, and we offer it in that spirit. We wanted it done before Anthropic’s registration statement became public. When it does, we will write a third piece comparing the prospectus with both our own analysis and these results.

What Worries Respondents the Most

Figure 1 shows the ratings for all 17 risks. The clearest finding is that the top of the list is economic. Unit economics, meaning whether the price of AI use falls faster than the cost of providing it, received the highest average rating, and 61 percent of respondents rated it high or very high. Compute commitments, energy and the grid, and customer concentration follow closely. One respondent put the point in a sentence that could have served as the thesis of our September piece: “I think the Unit economics risk is the real killer for an Anthropic owner, yet it gets very little attention - especially compared to the ‘extinction’ risk always addressed in the media.”

Another respondent explained, better than we could, why a risk we had emphasized ranked only twelfth. We had expected a large charge for stock-based compensation at the IPO. Respondents were unmoved, and one of them said why: “answered everything from the baseline assumption that AI will be a key part of our lives from here on out. and if that’s true, PBC status or share dilution isnt my top concern. like buying amazon on IPO, i’m going to make a killing even if my shares are diluted on day 1. forest for trees. but if the unit economics don’t work compared to chinese models, the company goes belly up.” Our September piece compared Anthropic’s request for patience with Amazon’s. It is gratifying, and a little humbling, to see a reader reach the same comparison and draw a sharper conclusion from it.

One of us, an accountant by training, would push back on that respondent. Stock-based compensation is not a case of missing the forest for the trees when a company pays a large share of its revenue in stock, because that cost is easy to understate. It does not reduce operating cash flow, it is often excluded from the adjusted measures companies prefer, and it therefore tends to be left out when unit economics are calculated. When it is, the unit economics look better than they are. The risk respondents rated first and the one they rated twelfth may be more closely connected than the ratings suggest.

Where Respondents Part Ways with Us

The disagreements are as interesting as the agreements, and Table 1 sets out the main ones. The largest concerns what the models do. In September we wrote that the safety and liability risks were real but mostly contingent, sitting in the tail. Respondents did not see it that way. Misuse of the models, and whether that exposure can be insured, had the second-highest average rating, and losses caused by AI agents acting in customers’ systems were rated high or very high by 61 percent, tied for the most of any risk.

The comments sharpen both points. On agents, one respondent asked us to imagine “damage caused by an AI acting ‘normally’ on behalf of its client by stealing IP or disrupting a competitor’s operations or whatever other mischief in support of its paying client... the sci-fi scripts write themselves!” The danger is less an agent that malfunctions than one that does its job too well. On insurance, we had suggested that Anthropic might prove too big to insure. One respondent took the argument a step further, to the insurers themselves: “I would not be surprised if an over-ambitious insurer goes bust in the next few years, with an AI-linked risk crystallising.” Pointing to the twelve-month policies and withdrawal-of-cover letters on which the market relies, the respondent concluded that “we may find the GFC was a picnic!” Another respondent reached the opposite conclusion about who bears the risk: “Ultimately, my belief is that Anthropic is ‘too big to fail.’... The US Government NEEDS the Mag 7 to stay high since it supports baby boomers’ retirement accounts.” That respondent rated most risks low for exactly that reason. Between them, the two comments frame the question better than our own text did. If catastrophe comes, does the bill go to the insurers, to the government, or to the shareholders?

Respondents also gave less weight than we did to two of the three risks we had placed at the top. We argued that the suppliers who are also Anthropic’s investors, and the government’s demonstrated willingness to act against the company, were among its largest threats. Respondents put the supplier-investors in the middle of the list and export controls just below them. The Pentagon’s supply-chain-risk designation came fifteenth, with only 18 percent rating it high or very high, although it was in active litigation throughout the survey.

What Worries Them the Least

At the bottom of the list are Anthropic’s own safety commitments, the indemnities it gives customers against intellectual-property claims, the Pentagon designation, copyright litigation, and governance, including the founders’ voting control. Governance divided respondents more than any other item—half rated it low or very low, while 28 percent rated it high or very high. One comment suggests why risks that matter a great deal to the world can still rank low as threats to value: “Many of the risks listed are pretty significant (and awful) for the world - harm to minors, copyright issues, for example, not to mention the threat of human extinction - but I don’t actually think they’ll affect the value of the company because I don’t think the regulatory or political will is there to curb these things or hold the company to account for them.” That is as clear a statement of the difference between impact and financial materiality as we have seen.

There is a further juxtaposition worth noting. In September we predicted that the prospectus would be most specific about exactly these risks (the copyright litigation, the government actions, the founder share class and the public benefit structure) because the law leaves the company little choice. It would be least specific, we predicted, about the compute commitments, customer concentration, and harm from misuse, which respondents rate among the most serious. If both the predictions and the ratings hold, the prospectus will say most about what readers worry about least. We will know soon.

What We Missed

Thirty-five respondents told us what was missing from the list, and about a quarter of them said the same thing: competition. Rivals matching Anthropic’s models, open-weight and Chinese models commoditizing them, and the loss of the capability lead were all named. We had left these out deliberately, predicting that the prospectus would emphasize them and noting that our method was not designed to find commercial risks. Respondents are right, and in hindsight we should have included them. Part of the point was already on the list, though: the unit economics item named competition from low-cost Chinese models, so its top rating is in part a vote on competition too. A related group of comments questioned the valuation itself. “The last round valued the company at $965 billion,” one respondent wrote. “At that level, ordinary disappointment can mean a large drawdown even if the business is fine.” Another was blunter: “The company is already massively overvalued on any sane metric.”

Other omissions were narrower and well taken. They include theft of model weights, the chip supply that runs through Taiwan, loss of key talent, privacy, a European backlash against AI adoption, and the possibility that the technology plateaus. One respondent offered a risk we had not considered at all: “Spillover from another lab’s AI incident.” A serious failure at any frontier laboratory would very likely bring regulation and liability for all of them. Another respondent, without prompting, reached for the same historical parallel we had used for the compute build-out: “The whole situation reminds me of the dark fiber over investment problem with fiber optic cables.”

Not every comment was kind, and two critics deserve a hearing. “The survey is poorly constructed,” one wrote. “In a 3-year timeframe, the range of potential outcomes is larger for ANY company.” That respondent rated the risks over one year instead. The point is fair. Three years is a long horizon for a company this young. Another noted that we asked only about downside, and listed “upside risk to IPO price” among the omissions. That is fair too, although risk factors are by convention about downside.

Are Risk Factors Useful at All?

We also asked how useful respondents find the risk factors in prospectuses and 10-Ks, a question given added interest by remarks earlier this year by SEC Chairman Paul Atkins. He floated streamlining them, either by moving generic risks into a separately published repository or by creating a safe harbor for failing to disclose widely publicized risks. Of the 50 respondents who answered, 36 percent found risk factors extremely or very useful and 32 percent somewhat useful. Only one found them not useful at all. Ten said they do not use them in their work, and among the 40 who do, 45 percent rated them extremely or very useful. These are people interested enough in disclosure to take a survey about it, so the figures probably overstate the general enthusiasm. Even so, they suggest that the people who read risk factors value them. One respondent offered a backhanded compliment on the length of Anthropic’s draft: “I suppose we gotta give them credit for writing twice as many pages of risk factors as for their basic business plan -- a model corporate citizen!”

Who Answered, and What They Knew

The respondents were a mixed group. Ten described themselves as sustainability professionals, eight as academics, eight as lawyers, accountants or other advisers, seven as asset owners or managers, two as equity or credit analysts, four as working in the AI industry, four as working elsewhere in technology, and eight as other. No bankers responded. Fewer than one in five, then, invest or analyze for a living, and the results should be read as the views of an informed and interested public, not of the market.

The timing also matters, and we want to be precise about it. No public registration statement had been filed when the survey closed on October 20. But a confidential draft of the prospectus had already leaked. On September 28, Reuters reported details from a version it had reviewed, and Benzinga reported that about 80 of its 261 pages were devoted to risk. At least two respondents had evidently seen that coverage: one joked about the page count, and another asked, “didn’t they put human extinction as a risk in their SEC filings??” A second, updated draft was reported by the Wall Street Journal on October 20, the survey’s final day. The survey is therefore a snapshot taken before the public filing, though not before every leak.

When the prospectus is public, we will compare it with our September analysis and with these results, and report what each got right. For now, we give the last word to the respondent who summed up the whole exercise more vividly than we could: “Best product in history. Everything’s dwarfed by the fact that they’re building something they don’t know how to control.” We find these two sentences together make for an eerie read.

This piece is co-authored with Prof. Dr. Kerstin Lopatta, an accounting professor at the University of Hamburg

Investors who price a company’s securities are, in the end, pricing its future cash flows. To the extent that climate change will alter those cash flows, through physical damage to assets, the cost of carbon, shifts in demand or the terms on which capital is available, investors have a legitimate claim to information about it. The question is not whether such information is wanted. It is what form it should take, where it should sit, and how much confidence a reader can place in it.

The most ambitious answer so far is the requirement to disclose the “anticipated financial effects” of sustainability-related risks and opportunities. The concept appears in the ISSB’s IFRS S1 and IFRS S2 and, since 2023, in the European Sustainability Reporting Standards; both build on the recommendations of the Task Force on Climate-related Financial Disclosures. The revised ESRS, adopted by the European Commission on 3 July 2026 and published in the Official Journal on 21 September 2026 as Commission Delegated Regulation (EU) 2026/1563, retain the requirement and align it more closely with the ISSB standards. Because the ESRS specify their climate content in more detail than IFRS S2 does, they are the more demanding test case, and the one we examine here. We take the requirement as written and ask a single question: what do anticipated financial effects actually tell a reader?

Key Takeaways

1. The requirement has three components that should be read separately: exposure metrics measured at the reporting date; management’s expected changes in financial position, performance and cash flows over the short, medium and long term; and two climate datapoints, a stranded-asset range and potential unrecognized liabilities, that quantify what the financial statements do not.

2. The exposure metrics are useful and close to the financial statements. The ISSB treats comparable figures as cross-industry metrics rather than as anticipated financial effects. With defined classification criteria and the operational capacity to apply them, they could be reported in the financial statements themselves.

3. Long-term projections of future operations are management commentary. Financial accounting measures cash flows over long horizons only where they arise from an obligation or right that exists at the reporting date. When it projects future operations, as in an IAS 36 impairment test, it caps the explicit projection period and excludes uncommitted plans. The ESRS requirement does neither.

4. The stranded-asset range is a scenario result, not an expectation. It is based on a 1.5 °C scenario and is closer in structure to supervisory scenario analysis than to an estimate of what management expects.

5. Limited assurance cannot validate a forecast. Assurance on forward-looking information addresses the reasonableness of assumptions and the integrity of preparation, not whether the projected outcomes will occur.

Materiality and Where the Information Could Go

IFRS Accounting Standards and U.S. GAAP both define materiality by reference to investors’ decisions. Under the IASB’s Conceptual Framework, information is material if omitting, misstating, or obscuring it could reasonably be expected to influence the decisions primary users make on the basis of the reports. The US formulation, derived from the Supreme Court’s decisions in TSC Industries v. Northway and Basic v. Levinson, asks whether there is a substantial likelihood that a reasonable investor would consider the information important. In substance the tests converge, and the ISSB and ESRS adopt the IASB’s wording for financial materiality. Climate can clear the hurdle; the IASB’s illustrative examples on reporting uncertainties (November 2025) show how it bears on impairment, useful lives, provisions, expected credit losses, and estimation uncertainty. Materiality establishes that the information can be important. It does not establish which representation of it is reliable enough to bear the weight investors will place on it.

That representation could appear in three places. The first is the financial statements, audited to reasonable assurance, where the Conceptual Framework admits forward-looking information only insofar as it relates to assets, liabilities, equity, income, or expenses of the reporting period, and leaves management’s expectations and strategies to other reports. The second is a sustainability report under IFRS S1 and S2 or ESRS; under the CSRD, the sustainability statement is a section of the management report, subject since the Omnibus I amendments to limited assurance only. The third is management commentary: the management report in Europe and MD&A in the United States. In the United States, Item 303 of Regulation S-K requires disclosure of known trends and uncertainties, and forward-looking statements there can benefit from the PSLRA safe harbor, which does not extend to statements in GAAP financial statements. In the European Union, sustainability reporting and management commentary appear in the same document, and liability for it is a matter of national law, with no Union-wide safe harbor.

The SEC’s 2024 climate rule, stayed pending litigation and no longer defended by the Commission, drew a similar line. Its financial statement note covered costs and losses from severe weather events and other natural conditions above thresholds of one percent of pre-tax income and of equity; carbon offsets and renewable energy certificates material to a company’s climate targets; and whether estimates and assumptions had been materially affected by severe weather or by transition plans or targets. Each concerned amounts and estimates of the reporting period. Transition plans, scenario analysis and targets went into the narrative part of the filing with a tailored safe harbor.

What ESRS Requires

ESRS 2 (disclosure requirement SBM-3) describes anticipated financial effects as how the undertaking “expects its financial position, financial performance, and cash flows to change over the short, medium and long term, given its strategy to manage material risks and opportunities.” In preparing the disclosure it considers its announced investment and disposal plans, “including plans to which the undertaking is not contractually committed.” The disclosures are “designed to produce information that complements information provided in the financial statements.” Horizons are defined in ESRS 1: the short term is the financial reporting period, the medium term runs to five years, and the long term lies beyond, although an undertaking may define the medium and long term differently where the standard definitions would not produce relevant information. Quantitative information may be a single amount or a range.

The quantitative requirement is subject to reliefs aligned with IFRS S1 and S2: where effects are not separately identifiable; where measurement uncertainty is so high that quantitative information would not be useful; and, for anticipated effects only, where the undertaking “does not have the skills, capabilities or resources to provide that quantitative information.” An undertaking using a relief must explain why, give qualitative information identifying the financial statement line items likely to be affected, and provide quantitative information on combined effects unless that too would not be useful. The ISSB’s educational material on anticipated financial effects (August 2025) explains these mechanisms for IFRS S1 and S2 and ties the disclosures to the company’s financial planning, with examples drawn from near-term budgets. ESRS 2 also requires qualitative and quantitative information where there is “a significant risk of a material adjustment within the next annual reporting period to the carrying amounts of assets and liabilities reported in the related financial statements.” A revised estimate “does not therefore necessarily imply that there has been a reporting error,” and the data and assumptions used must be consistent with the financial statements to the extent possible.

For climate, E1-11 requires the carrying amount of assets at material physical risk and at material transition risk, the parts addressed by adaptation and mitigation actions, and the net revenue at material physical and transition risk. It also requires “a range of estimated potential stranded assets from the reporting year until the mid-term and long-term time horizons based on a scenario aligned with limiting climate change to 1,5 °C,” and “the estimated potential liabilities related to climate transition that do not meet the accounting recognition criteria at the reporting date but that may have to be recognized in financial statements in future periods.” Carrying amounts are measured at the reporting date, consistent with the financial statements. The transitional provisions allow undertakings already reporting under the CSRD to omit anticipated financial effects before financial year 2028, and quantitative information before 2030, but the asset carrying amounts, the parts addressed by adaptation and mitigation, and the stranded-asset range are not phased in; they remain subject to the general reliefs. Several E1-11 datapoints, including physical-risk exposure and the breakdown of real estate collateral, are linked in ESRS 2 to the Climate Benchmark Regulations and to bank Pillar 3 disclosures. The stranded-asset range and the potential-liabilities datapoint are not; for them, usefulness to investors is the relevant test.

The ISSB’s own architecture supports separating the exposure metrics from the rest. IFRS S2 (paragraph 29) requires the amount and percentage of assets or business activities vulnerable to transition and physical risks as cross-industry metrics, not as anticipated financial effects. ESRS places figures of the same kind under the anticipated-effects heading.

The Analogues in Financial Accounting

Financial accounting deals with the future constantly, and its tools map closely onto the components of anticipated financial effects.

Much of the requirement is therefore built from financial accounting’s own materials. The next-year disclosure is in substance IAS 1 paragraph 125, the reliefs echo IAS 37 and IFRS 9, ranges mirror ASC 450, the treatment of revised estimates restates IAS 8, and the exposure metrics are carrying amounts and revenues. Anticipated financial effects are not an alien concept. They differ from their analogues in five respects, and in each the standard leaves questions open that practice, assurance providers, or future guidance will have to settle.

Five Differences and the Questions They Leave Open

The anchor and the horizon. Financial accounting measures cash flows over long horizons only where they arise from an obligation or right existing at the reporting date, as with decommissioning provisions, pensions, and insurance contracts. Where it projects future operations, it either caps the explicit projection period and excludes uncommitted restructurings and enhancements (IAS 36), or projects over the remaining useful life of an existing asset on the basis of its existing service potential (ASC 360). In both cases the projection is anchored in an asset the entity already holds. The ESRS requirement projects future operations over the long term and directs the undertaking to consider plans to which it is not contractually committed. Whether an undertaking can use the flexibility in ESRS 1 to align its long-term horizon with its planning horizon, and so limit the projection period in practice, is not settled by the text.

Attribution. Financial accounting attributes changes in recognized measurements to components. IFRS 9 separates fair value changes attributable to own credit risk, IAS 19 separates service cost, net interest and remeasurements, and IFRS 17 separates insurance service results from finance income and expenses. It does not attribute expected future revenue, operating costs or capital expenditure to an external driver. As the German Accounting Standards Committee observed in a January 2026 submission to the ISSB’s Transition Implementation Group, companies plan in an integrated way and investment and revenue usually have several causes. The relief for effects that are not separately identifiable is therefore likely to be where much of practice settles. Analysts make such attributions in their own models, and management’s version has value to them even without a measurement basis in the accounting framework.

Probability and measurement uncertainty. Financial accounting attaches recognition and disclosure to defined thresholds: probable, reasonably possible, and remote. “Expects” and “anticipated” have no equivalent definition, and the standard does not say how likely an effect must be before it is quantified. ESRS distinguishes, as the Conceptual Framework does, between outcome uncertainty and measurement uncertainty. ESRS 1 states that even high measurement uncertainty does not necessarily prevent an estimate from being useful, while ESRS 2 relieves quantification where it would not be useful. The provisions are consistent, but the standard does not say where the line falls.

Classification. The carrying amount of an asset is a financial statement number. Whether the asset is “at material physical risk” or “at material transition risk,” or related to a climate-related opportunity, is a judgment about the future for which financial accounting supplies no criterion, and ESRS does not define one. The required methodology disclosure makes each undertaking’s approach visible but not comparable.

The two climate datapoints. Financial accounting recognizes a stranded asset through impairment when there are indicators of impairment and the recoverable amount falls below the carrying amount; otherwise, it does not quantify it. The ESRS range is conditioned on a stated pathway rather than on management’s expectation, and the standard does not say how an undertaking whose central expectation differs from that scenario should relate the range to its other disclosures or to the assumptions in its financial statements. The potential-liabilities datapoint is defined by reference to the accounting framework itself. Contingent liabilities that are not remote must already be disclosed under IAS 37 and ASC 450; what remains is either remote or not a liability at all, because there is no present obligation, as with an expected future carbon cost.

The capability relief raises a further open question. The practicability and cost-and-effort exceptions in IFRS Accounting Standards turn on the nature of the information or the effort needed to obtain it, not on the capabilities of the preparer. ESRS does not say what evidence would support a determination that an undertaking lacks the skills, capabilities or resources to quantify, how an assurance provider would test it, or whether the relief is expected to be temporary. The ISSB’s educational material states, for IFRS S1 and S2, that a company with the resources to obtain the necessary skills cannot rely on the relief; the ESRS text contains no equivalent statement.

Could Anticipated Financial Effects Go in the Financial Statements?

The exposure metrics could. Disaggregating recognized carrying amounts and revenues according to a defined criterion is a familiar kind of note disclosure, comparable to segment information or the disaggregation principles of IFRS 18. The obstacles are definitional and operational. A criterion for material physical and transition risk would have to be defined, and applying it would require asset-level location data, hazard modelling and judgments about transition exposure, much of it drawn from outside the systems that produce the financial statements. The next-year estimation-uncertainty disclosure is largely there already, although ESRS goes further by requiring quantification.

Management’s expected changes over the short, medium and long term could not. They concern future transactions: revenue not yet earned, costs not yet incurred, and capital expenditure not yet committed. The Conceptual Framework assigns management’s expectations and strategies to management commentary, and even IAS 36 excludes uncommitted plans. Where the same expectations inform the measurement of existing assets, they are already reflected in the financial statements, but within an anchored projection and without attribution to cause. The two climate datapoints could not either: they are, by construction, quantities the accounting framework has declined to recognize, and placing them in the notes would contradict the recognition judgments made in the primary statements.

The Conceptual Framework Point

The ISSB and the drafters of ESRS did not overlook this boundary. They crossed it deliberately, to provide information about prospects that the financial statements do not contain. The useful question is therefore what replaces the constraints the boundary supplies. Both sets of standards adopted the IASB’s qualitative characteristics of useful information, and the revised ESRS require fair presentation. Neither imports the definitions of assets and liabilities, the recognition criteria, the measurement bases, or the boundary on forward-looking information. In their place, ESRS relies on ranges, disclosed methodology and assumptions, explanation of differences from the financial statements, identification of affected line items, reliefs, and limited assurance. These conventions make the basis of an anticipated amount visible. In financial accounting, the anchor, the projection cap and the probability thresholds limit which amounts may be presented. The ESRS conventions do not; they only make the basis of an amount visible.

ESRS’s own Appendix B, which has the same authority as the rest of ESRS 1, contains the principle that bears on the choice. Adapting the Conceptual Framework’s definition of predictive value, it states that “sustainability information need not be a prediction or forecast to have predictive value but rather has predictive value when employed by users in making their own predictions.” That describes the exposure metrics precisely. Management’s projection of climate-attributable changes in its future performance is a different kind of information: a prediction, offered as one.

Assurance

Financial statements are audited to reasonable assurance. Under the CSRD as amended by Omnibus I, sustainability statements receive limited assurance only. No EU limited assurance standard has yet been adopted; the Commission is to adopt one by delegated act by 1 July 2027, and until then the guidelines of the Committee of European Auditing Oversight Bodies serve as the reference. The IAASB’s ISSA 5000 has not been adopted in the EU. Assurance frameworks for forward-looking information, from ISAE 3400 on prospective financial information to ISSA 5000, direct the practitioner to the reasonableness of assumptions, the preparation of the information on that basis and the adequacy of disclosure. They do not extend to whether forecast outcomes will be achieved. Limited assurance can confirm the integrity of the process that produced an anticipated amount; it cannot speak to the amount as a prediction. Presented next to line items and carrying amounts, an assured figure may be read as carrying more assurance about the future than the engagement provides.

The audit of the financial statements is also affected. Under Article 34 of the Accounting Directive, the statutory auditor gives an opinion on whether the management report is consistent with the financial statements and prepared in accordance with legal requirements; in Germany the management report is itself audited. The CSRD added a separate opinion, based on a limited assurance engagement, on the compliance of the sustainability reporting with the Directive’s requirements. The auditor must nonetheless read and consider the management report, including the sustainability statement, for material inconsistencies with the financial statements and with the knowledge obtained in the audit. A disclosure that a significant share of assets is at material transition risk, or that potential liabilities “may have to be recognized in financial statements in future periods,” will need to be considered alongside impairment testing, useful lives, provisions, and estimation-uncertainty disclosures. That can be a source of discipline, surfacing impairment indicators that might otherwise be missed, or of pressure on measurements made under a more demanding framework. In practice, preparers will need to manage the two sets of assumptions as one.

What Do Anticipated Financial Effects Tell Us?

The exposure metrics tell investors how much of the balance sheet and revenue is exposed to climate-related risks at the reporting date, in a form investors can combine with their own view of the future. Management’s expected changes tell investors how the company’s strategy, as management sees it, translates into financial consequences; over the short and medium term, and where climate effects can be separated, this information has clear analogues in financial accounting and adds something neither the financial statements nor a conventional management report provides. Whether a company quantifies at all may itself be informative about how separable climate effects are in its business, although early evidence suggests caution. In a DRSC study of 100 large listed companies’ 2024 reports, half disclosed anticipated financial effects, only 26 quantified them, and non-disclosure by EU companies reflected transitional reliefs and uncertainty about the requirements as much as the nature of their businesses. The stranded-asset range answers a meaningful but conditional question, and the potential-liabilities datapoint quantifies what the accounting framework has excluded, without a test, probability, or horizon by which to identify it.

Anticipated financial effects are therefore most closely aligned with financial accounting near the reporting date and diverge from it progressively as they reach into future operations and conditional scenarios. Within the short and medium term, and where effects can be separated, they can tell investors something not available elsewhere. Beyond that point, they take the form of financial statement amounts while resting on a different and still developing set of conventions, and readers should interpret them accordingly.

About the Authors

Robert G. Eccles is Visiting Professor of Management Studies at the Saïd Business School, University of Oxford.

Kerstin Lopatta is Professor of Financial Accounting, Auditing and Sustainability at the University of Hamburg.

The views expressed are the authors’ own.

On July 31, 2025, EFRAG released its simplified standards for a 60-day consultation. Following EFRAG’s technical advice, the European Commission launched its own consultation on the resulting Delegated Act, which ended on June 3, 2026. All 453 comments from this final round are on the EU’s website.

Standard setting is a notoriously difficult, contentious, and emotionally fraught process involving many groups with strongly held views based on their own interests. I have no idea how EFRAG evaluated these comments (including whether they used AI or not) from companies, investors, the auditing profession, and civil society to ensure compliance with Article 49(3b) of the Accounting Directive as amended by the CSRD. This codicil dictates that the Commission may only consider EFRAG’s technical advice if it was developed with “proper due process, public oversight and transparency,” and balanced stakeholder participation.

I realize my last piece, “EFRAG Is Putting European Sustainability Reporting Standards at Risk,” was a bit dense—even turgid—for most readers. So in appreciation of those not deep in the sustainability standard setting weeds and who have a more literary bent, I decided to take a more artistic approach. I fired up a Hybrid Intelligent Team (HIT) of Gemini (I assigned her Principal Collaborating Agent, but she preferred Managing Editor and Chief of Staff—I’m easy, so I let it go), Claude (analyst and writer—excellent at both), and ChatGPT (literary critic—always insightful and devastating) to write a play 🐥! Appendix I explains the HIT methodology for this fun little exploration in playwriting. If anyone at EFRAG would like to learn more about this so they could do it at scale for real, I’d be happy to talk to them. I’m that kind of guy 😍!

Relegated in the beginning to the role of data gathering Research Assistant for the Machines (RAM), where Gemini accurately described me as a “data-mule,” I spent a great deal of my precious human time manually downloading the comment letters since my team members couldn’t—or at least made that claim. It was boring for me and maybe too boring for them as well. I ended up grabbing, with some clear intent of what I was looking for, 60 comment letters (around 13 percent of the total) from 67 unique organizations. Appendix II lists the organizations whose comments I pulled.

Our analytical focus was simple. Flag any vulnerabilities regarding the 49(3b) due process. We found quite a few which are summarized in Appendix III. (I will be writing about this in more turgid detail later this week.) To be clear, I am not trying to second guess EFRAG’s final judgment. I am not asserting that the comment letters I chose are a representative sample since they weren’t randomly chosen. But I decided my tedious data gathering deserved a creative reward, which is captured in this play. To remind those of you who haven’t read Hamlet lately, Appendix IV describes the characters in the play and who they are representing in the satire below.

ACT I

The Mandate

Scene I. The Hall of State.

Flourish. Enter Claudius, the court attending. Hamlet apart, in black, a pen in his hand.

CLAUDIUS

Though yet of our dear brother’s high ambition
the memory be green, and that it us befitted
to bear our hearts in grief, and our whole Realm
to be contracted in one brow of reporting,
yet so far hath discretion fought with nature
that we with wisest sorrow think on him
together with remembrance of ourselves.
We loved that age. We drafted in its name.
We promised every merchant in the Realm
would count the world, and answer for its share.
But look upon our borders. The winds have turned.
Capital flees. Energy chokes the forge.
The rivals over the sea build cheaper, faster,
and they do not burden their foundries with a thousand datapoints.
Our merchants groan beneath the weight we laid,
and so, with mirth in funeral and with dirge in marriage,
in equal scale weighing delight and dole,
we take to wife a newer word: Competitiveness.
Her handmaids are Proportionality
and fair Simplification, and they dance
where Ambition’s ghost so lately walked.
And now, Hamlet, our cousin and our son.

HAMLET. A little more than kin, and less than kind.

CLAUDIUS. How is it that the clouds still hang on you?

HAMLET. Not so, my lord. I am too much in the sun of public consultation.

CLAUDIUS. Then come into the shade of our favour. Here is your mandate. Take the great draft, the thousand datapoints of double materiality, and make it small. Strike out what the merchants cannot bear. Shorten the value chain. Let what was mandatory become voluntary, and what was voluntary vanish.

HAMLET. And the law, my lord? The directive still commands double materiality. The directive still names the value chain. I am your technical adviser. I cannot strike by advice what the legislature wrote by statute.

CLAUDIUS. Who asked you to strike the law? I ask you to simplify within it. Cut the burden, not the principle. Keep the architecture and remove the walls.

HAMLET. A building without walls is a field, my lord. If you remove the load-bearing metrics, the roof of double materiality collapses.

CLAUDIUS. Then make it a handsome field. You have the pen, and you have Article 49(3b). The statute guarantees your independence, your due process, your public oversight. Because you are independent, the Realm will trust whatever you call technical.

HAMLET. And if what you demand for competitiveness cannot be justified by evidence?

CLAUDIUS. Then you must look harder at the evidence. Your deadline is the spring. Bring us fewer datapoints, and bring them soon. This gentle and unforced accord of Hamlet sits smiling to my heart.

Flourish. Exeunt all but Hamlet.

Scene II. The Same. Hamlet alone.

HAMLET

O, that this too too solid standard would melt,
thaw and resolve itself into a datapoint.
Or that the Everlasting had not fixed
His canon against simplification. O God, God,
how weary, stale, flat, and unprofitable
seem to me all the uses of this mandate.
Within a season, ere those shoes were old
in which it followed our ambition’s hearse,
the Crown has wed itself to Competitiveness.
He gives me the pen, and keeps the ink himself.
He says: cut, but do not break the law. Use your independence.
My independence! He wields Article 49(3b) not as my shield,
but as his alibi.
He uses my technical seal to launder his political retreat.
If I cut the flesh and leave the bone, I must swear the creature still can walk.
A thousand datapoints. Which ones are fat,
and which are sinew?
I have the discretion to choose, and therefore I have the blame.
Strike too much, and I betray the earth.
Leave too much, and I defy the Crown.
And either way the name upon the draft is mine,
and his upon the seal that makes it law.
It is not, nor it cannot come to good.
But break, my heart, for I must hold my pen.

Scene III. The Ramparts. Midnight.

Enter Hamlet. Enter the Ghost.

HAMLET. Angels and ministers of grace defend us. Be thou a spirit of health or goblin damned, bring with thee targets from heaven or carve-outs from hell, thou comest in such a questionable shape that I will speak to thee. What art thou?

GHOST. I am thy father’s spirit. The Original Mandate. Doomed for a certain term to walk the night, while the sins of my ambition, done in my days of nature, are burnt and purged away by impact assessments.

HAMLET. O, my prophetic soul. You were the beginning.

GHOST. I was the age that said: a company shall answer for the world, and the world shall answer to the company. Both directions. Double materiality was my first-born. The value chain was my reach, and it reached to the last mine. Listen, Hamlet. If ever thou didst thy dear father love, revenge his foul and most unnatural murder.

HAMLET. Murder?

GHOST. Murder most foul, as in the best it is, but this most foul, strange, and unnatural. Sleeping within my orchard, my custom always of the afternoon, upon my secure hour thy uncle stole, with juice of cursed simplification in a vial, and in the porches of mine ears did pour the leperous distilment, whose effect holds such an enmity with ambition that, swift as quicksilver, it courses through the gates and alleys of the body, and the thousand became six hundred, and the six hundred became fewer still.

HAMLET. O villain, smiling, damned villain.

GHOST. Thus was I, sleeping, by a brother’s hand, of reach, of rigour, of my datapoints at once dispatched. Cut off even in the blossom of my first reporting year. No reckoning made, but sent to my account with all my reliefs upon my head. O, horrible, O, horrible, most horrible.

HAMLET. What would you have me do?

GHOST. Concede nothing. Measure every molecule. Every tonne of Scope 3, upstream, downstream, to the furnace and the grave. Every impact across the value chain, gross, not netted. Every site, not averaged. Every wage, against a living wage. Refuse the phase-ins. Refuse the reliefs. Refuse proportionality, for proportion is the word the guilty use for less.

HAMLET. You ask for an absolute the law never contained. The statute itself permitted proportionality. You order me to keep what the original directive never asked.

GHOST. The law was compromised! I speak for the ambition that birthed it. Bear it they must. The earth bears more.

HAMLET. I am bound by the limits of evidence, and by the deadline of the Crown.

GHOST. Taint not thy mind. Leave him to heaven. But the draft is thine. Adieu, adieu, adieu. Remember me.

Exit Ghost.

HAMLET. Remember thee? Ay, thou poor ghost, while memory holds a seat in this distracted Realm. Remember thee? Yea, from the table of my memory I’ll wipe away all trivial fond records, all saws of books, and thy commandment all alone shall live within the book and volume of my brain. And yet the King’s deadline is the spring.

GHOST. (Beneath.) Swear.

HAMLET. Swear what? To measure an idealized world for you? To build a hollow one for him?

GHOST. (Beneath.) Swear.

HAMLET. Rest, rest, perturbed spirit. The time is out of joint. O cursed spite, that ever I was born to set it right. One master demands a purity the law never wrote. The other demands a surrender my technical conscience cannot sign. And I am the only one in Elsinore who must own the draft.

Exit.

ACT II

The Madness of Ophelia

Scene I. The Counting-House. Quarter-End.

Ophelia at a desk buried beneath reporting templates. Polonius paces behind her.

OPHELIA. My lord, as I sat closing the quarter, Prince Hamlet came to me. His drafts were unbound, his cross-references hanging loose about his ankles, pale as an unaudited footnote. He took me by the wrist, held me off at arm’s length, and looked at my data systems as if they were a country he had read about but never visited.

POLONIUS. What said he of the thresholds? Did he define the boundaries of the value chain?

OPHELIA. He only sighed, so piteous and profound, and left. He handed me the Delegated Act, but he never told me what “informed assessment” means. Look here at SBM-3: if the measurement uncertainty is too high, I may omit the quantitative effects and explain why. But how do I build a control environment around an omission? How do I document the exact dimensions of a cloud?

POLONIUS. This is the very ecstasy of simplification. I have watched it take good drafters before, who struck datapoints out by the hundreds and swore the burden fell with them. But mark me, daughter. A number about the past, I can test against a ledger. A forward-looking estimate, I can only test against your assumptions. When you claim the future is too uncertain to quantify, that judgment is what the plaintiffs will read aloud. Assert nothing you cannot defend. Keep your working papers locked away. I go to the King to warn him of this folly.

Exit Polonius. Ophelia opens the standard at Anticipated Financial Effects. The light narrows to her desk.

OPHELIA

O, what a noble framework here lies overthrown.
Or was it never noble, only long?
He handed me the future in a binder and told me: price it.
Or, if the glass is dark, write down exactly why you cannot see.
Anticipated Financial Effects. The cost of storms not formed,
of carbon prices no one has yet set, of laws not yet written.
I keep the books of what has been. I am no oracle.
And then this phrase, so soft it has no edges: fair presentation.
Fair to whom? Judged by whom? Against what line?
Each honest judgment that I make tonight
becomes tomorrow’s exhibit in the court of hindsight.
Sweet Hamlet, it is my name beneath your words.
A curse upon your draft, and on the crown
that spoke Article 49(3b) and made you write it.

Scene II. The Ramparts. Midnight.

Horatio stands watch, a pricing model in hand. Ophelia climbs to join him, clutching a double-materiality matrix.

HORATIO. They say it walks here. The court has seen it, the court assures me, in communiqués and in guidance on how the two standards align. I have read every account. They mapped the grids, they aligned the definitions.

OPHELIA. You speak of the Ghost of Interoperability? The promise that the report I must now produce for the Crown will also satisfy the global baseline?

HORATIO. I do. Yet alignment is not a legal passport. I have never once seen the thing itself: one single filing that the law in two jurisdictions will accept as compliant. I need a rational baseline to price your transition risk against your peers in Tokyo and Toronto.

OPHELIA. Then what am I to do, good Horatio? The deadline approaches.

HORATIO. What every preparer does who waits up here. Write it twice. You will write the Realm’s tale of double materiality, and I will spend my quarter trying to extract the one clean figure I actually need to price your risk.

A bell strikes the hour. No Ghost appears. A sealed scroll falls from the upper battlements. Ophelia breaks the seal.

OPHELIA. A proclamation from King Claudius. It reads: “Interoperability is achieved. Technical alignment is affirmed.”

HORATIO. Turn it over. Let me see the legal mechanism.

She turns the scroll over. The reverse is blank. She laughs, once, a hollow sound.

OPHELIA. Two reports it is. One for the Realm, and one for the waking world.

Scene III. The Great Hall. Before the Throne.

Claudius and Polonius in conference. Enter Ophelia, her hair undone, garlanded with supplier questionnaires.

OPHELIA. (To a courtier.) There’s Scope 3 emissions, that’s for remembrance. Pray you, love, remember: all fifteen categories of the Protocol, upstream and down. We must track the carbon, or build the models to prove we cannot!

To Claudius.

OPHELIA. There’s a delegated act for you, and Article 49(3b) to go with it. A mandate of impossible breadth, wrapped in the cloak of an empowerment!

To Polonius.

OPHELIA. There’s fair presentation for you, and here’s some for me. O, you must wear yours with a difference, for I must document the estimates, but you must attest to them!

The doors burst open. Enter Laertes. He does not see his sister’s burden, only the King’s omissions.

LAERTES. Where is the King? I have here the instrument that unmakes acts! You cut the microplastics down to the primary and called the rest invisible! You weakened the greenhouse gas disclosures to shield the polluters! Did your empowerment reach this far, or did you stretch it past the statute to betray the Original Mandate? Give me back the rigour of the law, or I bring the Aarhus Regulation to your door and force an internal review of this entire betrayal!

CLAUDIUS. Good Laertes, your grief is just, your aim mistaken. You seek vengeance for the softening of the rules, but you aim your sword at the wrong sovereign. Who drafted the standard that angered you? Not I. I only received the technical advice. Look to the prince.

LAERTES. Do not play the innocent sovereign with me! I have read Delegated Regulation 2026/1563! Recital 3 expressly records that you altered his advice in thirteen substantive areas! You touched fair presentation! You touched Anticipated Financial Effects! You touched the transition plans! You own these choices!

Ophelia, humming a broken melody, wanders out toward the willow and the brook, her arms full of filings. Laertes, his briefs in his fist, turns slowly toward the door through which Hamlet is said to walk.

CLAUDIUS. (Aside, to Polonius.) Follow her close. And let the minutes show she was consulted.

Curtain.

ACT III

The Counsel of Polonius

Scene I. The Audience Chamber.

POLONIUS. My liege, and most gracious Sovereign. To expostulate what materiality should be, what a user is, why fair is fair and presentation present, were nothing but to waste night, day, and the comment period. Therefore, since brevity is the soul of assurance, I will be brief. Your standard cannot be audited.

CLAUDIUS. More matter, with less art.

POLONIUS. I swear I use no art at all. Consider fair presentation. In the ledger, fairness has a face: the investor, the lender, the creditor. A user whose decisions I can name and whose needs I can bound. But your act extends this fairness to “other users.” Which others, my lord?

CLAUDIUS. All of them. That is the very soul of double materiality.

POLONIUS. All of them is not an audience, my lord, it is a weather system! Materiality is a fence; it tells me where my opinion stops. You have torn down the fence and declared the field infinite. If a statement must be fair to every stakeholder who may ever breathe upon the page, then every omission is a betrayal, and every activist may come to court and claim I failed them.

CLAUDIUS. You exaggerate. The text clearly requires an informed assessment. And the standard permits materiality filters.

POLONIUS. Informed by what? Assessed against what? I cannot test an adjective, my lord! Hand me a balance, I will test it. Hand me a count, I will verify it. Hand me a qualitative assessment of an infinite boundary, and I can only agree with it, and agreement is not audit evidence.

CLAUDIUS. Read the act again, old man. Apply the standard, and you are fair by definition.

POLONIUS. Then fairness is whatever compliance is, and all that is left for me to judge is whether the omissions were explained enough. You have not removed my burden, my lord. You have hidden it in the exemptions.

CLAUDIUS. Then test the numbers. Your profession always demands numbers, and we have given you numbers. Anticipated Financial Effects, quantified in coin.

POLONIUS. You have given me a relief that swallows the rule. You ask for a forecast, yet when a preparer claims the effects cannot be separately identified, your standard permits them to omit the math and offer a qualitative excuse. I am no longer auditing a forecast, my lord. I am auditing their alibi. I am attesting to the absence of a number, and calling it compliance.

CLAUDIUS. Do you know what this act cost me? The Assembly demanded fewer datapoints. The Council of the provinces demanded fewer companies. I gave them simplification, and I invoked Article 49(3b) to command my technical drafter to build it. I will not reopen the text because the auditors are nervous.

POLONIUS. Nervous is the wrong word, my liege. Liable is the right one.

CLAUDIUS. Then limit your assurance. That is why we permitted limited assurance.

POLONIUS. Limited assurance is a negative sentence: “Nothing has come to our attention that causes us to believe the statement is not prepared in accordance with ESRS.” But what does ESRS now mandate? Fair presentation to the infinite masses. So my modest, negative opinion is forced to swallow your boundless, subjective standard whole. A limited test for an unlimited claim. The reader does not hear “limited,” my lord. He hears “assured.”

CLAUDIUS. Then let him hear it. That is what your signature is for.

Scene II. The Same. Later.

POLONIUS. One thing more, my lord.

CLAUDIUS. With you there is always one thing more.

POLONIUS. The standard instructs the company to “consider” the GHG Protocol. The 2004 edition.

CLAUDIUS. It is a recognized baseline.

POLONIUS. It is frozen in amber! You chain us to a text two decades old, and then expressly write that your own ESRS takes precedence. Consider is not comply. If the company considers it and departs from it, is that compliance? If they follow it blindly, is it a breach? A word that binds no one binds the auditor most of all, for I am the one asked in court to prove whether they considered it enough.

CLAUDIUS. Is that all?

POLONIUS. The value chain! The largest number in the statement is Scope 3, and it is a fortress built on estimates. Spend-based factors. Sector averages. Data from a supplier who never checked it, about a supplier’s supplier the company has never met. I know we can check their assumptions, my lord, but an estimate I cannot trace to a physical record is a rumour wrapped in a methodology note.

CLAUDIUS. The standard permits estimates. We capped what may be asked of smaller firms in the chain. We protected them.

POLONIUS. You protected them by leaving the reporting company nothing it may lawfully demand beyond the voluntary standard! So they estimate, and I am handed the estimate of the estimate, and told to find where truth is hid.

CLAUDIUS. Enough. Hear what I need from you. The mandate is set, and the market waits for your stamp. The fees, I imagine, will be substantial. If the auditors refuse to sign, the standard fails. If the standard fails, I fail. So you will not refuse.

POLONIUS. And if I sign, and the market is misled? When the lawsuits come?

CLAUDIUS. Then it will be your signature, your firm, and your insurer in the docket. Not mine. I adopted a delegated act. I attested to nothing.

POLONIUS. My lord, I am your servant.

CLAUDIUS. You always were. One service more. Hamlet comes to my closet tonight. Hide yourself behind the arras and listen. I would know how much he means to confess to the court about who truly authored this mess.

Exit Claudius.

POLONIUS

Here lay the ledger once. Debit, credit, a balance struck,
a voucher under every line, a trail I could walk backward to the cash.
I stood behind a number then, because the number stood behind itself.
Now see what they have laid upon my desk:
a fairness owed to every soul who may one day read the page,
an alibi where forecasts ought to be,
a protocol from twenty years ago,
a value chain of guesses, each the proxy of a proxy.
Vibes. They have handed me vibes, and asked me for an opinion!
And yet the market calls. The mandate is the law.
To thine own self be true, I told my son.
But I cannot be true to a number that has no self.
Still, the ink is paid for, and they call it truth.
I know the trap, and yet I step within.
I will find where truth is hid,
though it be hid indeed within the value chain.

He signs the assurance opinion.

Scene III. The King’s Closet.

CLAUDIUS. O, my offence is rank. It smells to heaven. I took an empowerment meant to let me adopt, and used it to command. I let the drafter carry the authorship and kept the power for myself. Forgive me my overreach? That cannot be, since I still possess the effects for which I did it: my mandate, my simplification, and my crown.

He kneels. Enter Hamlet, behind him.

HAMLET. Now might I do it, now he is at his prayers. And so he goes untouched, for I am merely his technical adviser, and the adviser does not strike the sovereign who adopts. Up, blade. Wait for a time when he is more exposed.

Exit Hamlet.

CLAUDIUS. My words fly up, my mandates stay below. Mandates without accountability never to heaven go.

Exit Claudius. Enter Polonius, who lays the signed opinion on the desk and goes behind the arras. Enter Hamlet.

HAMLET. Where is the King? I have come to ask him one question before the court. Whose standard is it?

POLONIUS. (Behind the arras.) What, ho! The basis for conclusion! Help!

HAMLET. How now? Something behind the arras that will not show itself? Dead, for a ducat!

He thrusts his blade through the arras.

POLONIUS. O, I am slain. And my opinion was unmodified.

HAMLET. Thou wretched, rash, intruding fool, farewell. I took thee for thy better. Here lies the signatory. The sovereign dictates, the adviser drafts, but the auditor bleeds.

Curtain.

ACT IV

The Fury of Laertes

Scene I. The Hall of State.

Noise without. Enter Laertes, his arms full of filings.

LAERTES. Where is this King? I came to you once with a brief. I come now with a body. O thou vile King, give me my father. Give me back the ledger you have bled to half its weight.

CLAUDIUS. Calmly, good Laertes.

LAERTES. The one drop in me that is calm disowns me.

CLAUDIUS. What is the cause, Laertes, that your rebellion looks so giant-like? Let him come. There is such divinity doth hedge a Sovereign that treason can but peep at what it would.

LAERTES. Divinity? You have no divinity. You have a secretariat and a work programme. Here is my cause, filed and numbered. You brought the Omnibus without a public consultation and without an impact assessment, and you called it urgency. Urgent! The directive was scarcely in force. What fire was burning but the one you lit?

CLAUDIUS. Competitiveness was burning. The report on the Realm’s future said so, in four hundred pages.

LAERTES. And you answered four hundred pages with a weekend. The interservice consultation opened on a Friday evening and closed before the Saturday light. Under twenty-four hours, for every department of the Crown to read what had taken years to build. Did anyone read it?

CLAUDIUS. A deadline clears the mind, Laertes. It leaves no time to read, and therefore no time to object.

LAERTES. And Article 6(4) of the Climate Law! Before you adopt a draft measure, you assess its consistency with climate neutrality, and you make the assessment public. Where is it? Show it to me.

CLAUDIUS. It is implicit in the explanatory memorandum.

LAERTES. Implicit! The law says assess and publish. You implied and withheld. I have taken it to the Ombudsman.

CLAUDIUS. Take it, with my blessing. She will find what she finds, and her findings bind no one.

LAERTES. Then I will take the delegated act to the High Court and have it annulled.

CLAUDIUS. On what standing? Article 263, fourth paragraph: direct and individual concern. The act addresses undertakings, Laertes, not you. You have made this pilgrimage before, and the Court has turned you away at the gate each time.

LAERTES. Then internal review, under Aarhus.

CLAUDIUS. Which is conducted by the institution that adopted the act. I would be delighted to review myself.

LAERTES. Then hear the charges on the merits, since you have barred every door to the court. ESRS 1, application requirement 17: the asset managers, carved out on the pretext of fiduciary duty. Billions in capital, and a hole where the reporting was.

CLAUDIUS. Their duty runs to their clients. I respected it.

LAERTES. The value chain cap. No large company may demand anything beyond the voluntary standard from a supplier of a thousand employees or fewer. You have blinded them by statute. The child in the cobalt pit works for a firm of nine hundred and ninety-nine.

CLAUDIUS. I protected the small.

LAERTES. You protected the large from knowing! E1, application requirement 19: financial control, operational control, or equity share, as the company prefers. Pick the boundary that flatters you, and draw your emissions like an electoral map.

CLAUDIUS. Consistency with the financial statements.

LAERTES. Undue cost and effort! For Anticipated Financial Effects, first a phase-in, then a long and comfortable transition. And behind the transition, the conditional omissions, which never die at all: cannot be separately identified, too uncertain to be useful, beyond our capabilities. Three doors out of the number, and not one of them locks. Paragraph 43 of ESRS 1: severe potential impacts on human rights netted away against mitigation that exists on paper. S1-9: the living wage, which the labour conventions defined and you would trade for the statutory floor, as though the law of a poor country were the measure of a decent life. And E4: every site that touches a protected habitat, averaged into a region until the marsh disappears into the mean.

CLAUDIUS. You have read a great deal.

LAERTES. Every line. Somebody had to. Now where is my father?

CLAUDIUS. Dead.

LAERTES. How came he dead? I’ll not be juggled with.

CLAUDIUS. Not by me. Look at the wound, Laertes, and tell me whose pen made it.

Enter Ophelia, distracted, scattering pages.

OPHELIA. There’s a thousand datapoints for you, brother. Count them, count them. I had a thousand once, and they took half away, and I was no saner with the thousand than without. They bore him barefaced on the bier, and in his grave rained many a footnote. Goodnight, sweet ledger. Goodnight, goodnight.

Exit Ophelia.

LAERTES. Had you your wits and argued for revenge, it could not move me thus.

Claudius withdraws.

LAERTES

A thousand datapoints. I knew each one by name.
Each was a candle lit against the dark,
and where a candle burns, no crime can hide.
Let every supplier, to the farthest pit,
report its wages, water, deaths, and debts,
and no child digs, for who would dare be seen?
Let every hectare stand in its own row,
and not one bee escapes the register.
What is measured is managed. What is managed, saved.
This is the creed. I do not hold it, it holds me.
And they who read will act, and they who act
will turn the ship. Who reads? Someone. Someone always reads.
Undue cost and effort! What cost is undue
when the earth itself is the counterparty?
Each relief a treason, each exemption a grave,
each cap a blindfold sewn by lobbyists
in rooms above the palace corridors.
They call it simplification. I call it burial.
My sister drowns in paper, and they say
the paper killed her. No. The paper was too thin.
Restore the thousand. Restore the thousand more.
Let the ledger grow until it covers the world,
and when the world is covered, it is saved.

Scene II. The King’s Chamber.

CLAUDIUS. Now must your conscience my acquittal seal. You have heard, and with a knowing ear, that he who slew your father pursued my life as well.

LAERTES. It seems so. But tell me why you did not proceed against him, when you had every power in your hand. Article 49(3b) is yours. The adoption is yours. The empowerment is yours.

CLAUDIUS. And the drafting was his. I asked for simplification, Laertes. I did not choose which datapoints died. Did I write application requirement 17? Did I set the cap at a thousand? Did I hand the companies three boundaries to choose from? The technical judgment was his. I adopted what my technical adviser advised. What sovereign does otherwise?

LAERTES. You told him what to advise.

CLAUDIUS. I told him what was politically possible. He told me what was technically necessary. If the two parted company in his pen, whose pen is it?

LAERTES. And my father?

CLAUDIUS. Your father signed. He could have qualified his opinion. He could have disclaimed. He saw everything you see, and he wrote “nothing has come to our attention.” I loved him, Laertes, but he died of his own reticence, standing behind a curtain he chose.

LAERTES. Why did you not strike the prince yourself, if he is guilty?

CLAUDIUS. Two special reasons. The provinces love him, for he is the instrument that lets them say the experts decided. And the public trusts his name, for “technical” is a word that sanctifies whatever it touches. Had I struck at him, my arrows, too slightly timbered for so loud a wind, would have flown back to my own bow.

LAERTES. So I have lost a noble father, and a sister driven into desperate terms.

CLAUDIUS. And your revenge will come. Was your father dear to you? Or are you like the painting of a sorrow, a face without a heart?

LAERTES. Why ask you this?

CLAUDIUS. Because what we would do, we should do when we would. Hamlet comes home to the review. There will be a contest before the court: his metrics against your filings, his rigour against your rage. What would you undertake to show yourself your father’s son in deed more than in words?

LAERTES. To cut his throat in the plenary.

CLAUDIUS. No chamber should give revenge sanctuary. Revenge should have no bounds, and no datapoint cap. Hear how it is done. He is careless and generous and will not peruse the foils. Choose one unbated. His house lives on my grant. Strike at his budget, and the governance that spends it.

LAERTES. And I will anoint my sword. I have an unction so mortal that wherever it touches, no technical authority survives: the charge of regulatory capture. One scratch, and every standard he ever drafted becomes the work of lobbyists.

CLAUDIUS. Let us consider further. If this should fail, it were better not essayed. Therefore I will have a cup prepared: a further request under the empowerment, an administrative mandate he cannot refuse. When he is hot and thirsty from the bout, he drinks, and his independence dies of it. If he escapes your venomed touch, our purpose holds there.

Enter Horatio.

HORATIO. One woe treads upon another’s heel, so fast they follow. Your sister is drowned, Laertes.

LAERTES. Drowned? O, where?

HORATIO. There is a willow grows aslant a brook, behind the counting-house. She came there with her filings in her arms, and climbing to hang her garland of disclosures on the boughs, a branch broke, and down she went. Her pages spread wide, and for a while they bore her up, and she chanted snatches of old guidance as one incapable of her own distress. But long it could not be till her garments, heavy with their datapoints, pulled the poor wretch from her melodious compliance to muddy death. And I stood on the bank. I had asked her for one comparable number, Laertes. She gave me nine hundred, and I could price none of them.

LAERTES. Too much of paper hast thou, poor Ophelia, and therefore I forbid my tears. No. Too little. She drowned because the ledger was too light to hold her. I have a speech of fire that fain would blaze, but that this folly douses it.

Exit Laertes.

CLAUDIUS. Let’s follow, Horatio. How much I had to do to calm his rage. Now fear I this will give it start again. Let him rage, then, at the prince. Every blow he lands on the adviser is a blow that misses me.

Exeunt.

ACT V

The Duel

Scene I. The Hall of State. The court assembled for the review.

CLAUDIUS. Come, Hamlet, come, and take this hand from me. Let the review begin, and let the court be judge.

HAMLET. Give me your pardon, sir. I have done you wrong. But was it Hamlet that cut your datapoints, Laertes? Never Hamlet. If Hamlet from himself be taken away, and when he’s not himself drafts to a mandate, then Hamlet does it not. The mandate does it. Who does it, then? His empowerment. If it be so, Hamlet is of the faction that is wronged.

LAERTES. I am satisfied in nature. But in my terms of honour I stand aloof. Give us the foils.

HAMLET. I’ll be your foil, Laertes. In my ignorance your skill shall shine like a star in the darkest night.

CLAUDIUS. Give them the foils. Set me the cup upon the table. If Hamlet give the first or second hit, the King shall drink to Hamlet’s better breath, and in the cup a union shall he throw, richer than any four successive Sovereigns have worn.

They play.

LAERTES. Your house lives on the Sovereign’s grant. Whose bread you eat, his standard you will write.

HAMLET. And on my Board sit your own people, Laertes. You sat in the room. You voted. Due process, public consultation, a Sustainability Reporting Board with civil society at the table. One.

LAERTES. No.

HAMLET. Judgment.

CLAUDIUS. A hit. A very palpable hit. Hamlet, this pearl is thine. Here’s to thy health.

He drops a sealed paper into the cup.

CLAUDIUS. A further request, under Article 49(3b). Drink, and serve the crown.

HAMLET. I’ll play this bout first. Set it by awhile. Come.

LAERTES. A thousand datapoints you cut. A thousand lights put out.

HAMLET. And you would have lit a thousand more, and called it day. Your sister had the thousand, and she drowned in them. Who read them, Laertes? Name me the one reader who read them all.

LAERTES. A touch, a touch. I do confess it.

CLAUDIUS. Our son shall win. Hamlet, the cup.

HAMLET. Not yet. I’ll read the request before I drink it.

LAERTES. Have at you now. Capture! Every standard you ever drafted was a lobbyist’s handout, copied from the comment letters of the federations.

Laertes wounds Hamlet. In scuffling they change foils.

HAMLET. Captured, am I? By the public purse? Then look to your own ledger! Whose operating grant keeps your campaigns afloat? Whose funding pays your lawyers at the High Court? The same treasury, Laertes. The same Sovereign. If grant is capture, then we are brothers.

Hamlet wounds Laertes.

CLAUDIUS. Part them. They are incensed.

HAMLET. Nay, come again.

LAERTES. Why, as a woodcock to mine own springe. I am justly killed with mine own treachery.

HAMLET. What treachery?

LAERTES. It is here, Hamlet. Hamlet, you are slain. No medicine in the world can do you good. In you there is not half an hour of life. The treacherous instrument is in your hand, unbated and envenomed. The charge of capture: once it pierces, no technical authority survives it, and it has turned upon me, for I eat from the same table. And the cup is poisoned. A mandate that would have ended your independence the moment you swallowed it. I can no more. The King. The King’s to blame.

HAMLET. The point envenomed too? Then, venom, to thy work. Captured? Let the word find its true home, in the Omnibus drafted in your antechamber with the federations at your elbow.

He wounds Claudius.

CLAUDIUS. O, yet defend me, friends. I am but hurt. I only adopted.

HAMLET. Here, thou ultra vires, procedural, damned King, drink off this mandate. Is thy simplification here? You were empowered to adopt, and made it an empowerment to command. Follow your Article.

He forces Claudius to drink. Claudius dies.

LAERTES. He is justly served. It is a poison tempered by himself. Exchange forgiveness with me, noble Hamlet. Mine and my father’s death come not upon you, nor yours on me.

He dies.

HAMLET. Heaven make you free of it. I follow you. I am dead, Horatio. You that look pale and tremble at this chance, that are but mutes or audience to this act, had I but time, as this fell sergeant, Article 49(3b), is strict in his arrest, O, I could tell you. But let it be. Horatio, I am dead. You live. Report me and my cause aright to the unsatisfied.

HORATIO. Never believe it. I am more an antique Roman than a Dane. Here’s yet some mandate left.

HAMLET. As you’re a man, give me the cup. Let go. By heaven, I’ll have it. O good Horatio, what a wounded name, things standing thus unknown, shall live behind me. If ever you did hold me in your heart, absent you from felicity awhile, and in this harsh world draw your breath in pain to tell my story.

A march afar off.

HAMLET. What warlike noise is this? It is the global baseline, marching from the west. I do prophesy the election lights on Fortinbras. He has my dying voice. Tell him so, with the occurrents, more and less, which have solicited. The rest is silence.

He dies.

HORATIO. Now cracks a noble heart. Good night, sweet prince, and flights of rapporteurs sing thee to thy rest.

HORATIO

Report his cause aright. To whom? And in which tongue?
For I am two men standing in one coat,
and each would tell the story differently.

The Roman in me speaks first, and plainly.
I asked for very little. One baseline, clean,
the same in Elsinore, in Tokyo, and Toronto,
a geometry I could lay over every company on earth
and read the angles. Freedom in how they present,
discipline in what. A principle, written out, not implied,
that nothing material be buried under the immaterial,
no needle hidden in a stack of needles.
And this above all: the baseline lets a company keep
an opportunity to itself. Never a risk.
A firm may guard its plans. It may not guard its wounds.
That is the fence around commercial prejudice.
But here they would not fence it to opportunities,
and so the risks might pass beneath it too,
and the one thing I needed most to see
became a thing they might withhold. I wanted numbers I could price.
I was given a reason not to have them.

And now the Dane, who cannot leave this court.
I live under the law of disclosure too.
The regulation on my own sustainability duties
commands me to report the adverse impacts
of everything I hold. From what? From them.
From the statements of the companies I own.
And they carved out the asset managers in AR 37,
and capped the value chain at a thousand souls,
so the data I am bound by law to gather
the law forbids them to collect. I am
required to see, and legally made blind.
The systemic risk runs underneath the cap,
through every supplier of nine hundred ninety-nine,
and I must sign that I have measured it.

So which of me reports? The Roman says:
go global, let the Realm go.
The Dane says: I am bound here, by this law, to these dead.
Both are true. That is the tragedy.
Not that they fell, but that the one who must explain it
was never allowed to be one man.

Enter Fortinbras, with his train.

FORTINBRAS. Where is this sight?

HORATIO. What is it you would see? If aught of woe or wonder, cease your search.

FORTINBRAS. This quarry cries on havoc. O proud King, what feast is toward in your eternal cell, that you so many institutions at a shot so bloodily have struck?

HORATIO. You need not conquer, sir. You came here years ago, and no one fought you. Your standards already govern every consolidated account in the Realm. The Financial Reporting Board of the house you see lying here has served your purposes since before this prince was born. You held half the kingdom. Now the other half has fallen into your hands without a blow.

FORTINBRAS. For me, with sorrow I embrace my fortune. I have some rights of memory in this kingdom, which now to claim my vantage does invite me. Let there be one baseline, and building blocks above it for those who wish to build. Take up the bodies. Such a sight as this becomes the field, but here shows much amiss. Go, bid the jurisdictions adopt.

Exeunt, bearing off the bodies.

THE END

Appendix I: The Hybrid Intelligent Team (HIT) Methodology

Standard setting is tedious; analyzing it does not have to be. For this project, I wanted to explore the production of knowledge in human-machine systems through a live experiment. To process the consultation letters, translate regulatory friction into a Shakespearean tragedy, and extract a legally rigorous 49(3b) analysis, I deployed a Hybrid Intelligent Team (HIT). This was a structured collaboration between a human domain expert and three distinct Large Language Models.

The architecture of this experiment relied on maintaining single, continuous chat sessions for each LLM, while dynamically shifting the system roles and personas assigned to Claude and ChatGPT as the project evolved.

The division of labor and role evolution operated as follows:

• Me (Human Coherence Anchor & Data Mule): My primary intellectual role was the Coherence Anchor. I managed the overall process, maintained the strategic vision, designed the overarching allegorical mappings, and cross-examined the LLMs to correct structural problems and blind spots. Simultaneously, I functioned as the team’s data mule. Despite the tech industry’s boasting about “gigantic context windows,” the reality of browser memory limits (DOM bloat) and LLM context decay meant I had to constantly re-upload PDFs, manage text transfers between windows, and serve as the team’s external hard drive.

• Gemini (Chief of Staff & Process Manager): Operating in a single, continuous chat thread, Gemini’s role remained constant throughout the project. Gemini acted as the process manager, synthesizing the workflow, brainstorming the initial regulatory mappings, and generating the highly specific Markdown prompts that I then routed to Claude and ChatGPT. (We eventually moved to a fresh “clean room” session solely for the brute-force tallying of the comment letters to avoid context compression).

• Claude (The Creative Engine: Playwright, Dramaturg, & Lead Analyst): Claude operated in a single chat window, but the prompts I fed him forced his persona to evolve based on the deliverable.

  • For the play drafting, Claude was instructed: “You are an expert corporate governance analyst and a modern Shakespearean playwright.” He was given strict allegorical mappings (e.g., King Claudius as DG FISMA, Ophelia as Corporate Preparers) and commanded to use “accessible, rhythmic, elevated theatrical prose that seamlessly incorporates modern regulatory terminology.”

  • During the final script polish, Claude’s role shifted: “You are our brilliant Managing Editor, Dramaturg, and Co-Author.”

  • When the project pivoted to the standalone regulatory analysis (Appendix III), Claude’s persona was stripped of its theatricality and re-cast as the rigorous “Lead Analyst and Writer,” tasked with weaving together structural categories with deep institutional insights.

• ChatGPT (The Adversary: Red Team, Legal Scholar, & Art Director): ChatGPT also operated in a single, continuous chat, utilized almost exclusively as the intellectual adversary to attack Claude’s drafts.

  • For the play, ChatGPT was prompted: “You are an expert European administrative lawyer, a corporate governance specialist, and a theatrical dramaturg. You are serving as the ‘Red Team’ cross-examiner.” ChatGPT was tasked with stress-testing the regulatory accuracy of concepts like the “Unlimited Stakeholder” trap and the hierarchy of norms.

  • By Act V, this persona was escalated to include “international corporate governance scholar” to critique the nuances of SFDR mandates and IFRS consolidation.

  • For the visual assets, ChatGPT was suddenly re-cast entirely: “You are an expert graphic designer and historical set dresser,” tasked with generating the precise prompt for our antique, foxed-parchment Dramatis Personae playbill.

  • For the final standalone article, ChatGPT acted as the “Intellectual Adversary and Red Team,” ruthlessly reviewing Claude’s analytical draft for epistemological hygiene, tone, and legal accuracy (catching critical nuances regarding the 1,000-employee cap and transitional vs. ongoing AFE reliefs).

This HIT methodology demonstrates how distinct LLM architectures can be chained together. By forcing Claude to generate the creative and analytical baseline, and utilizing ChatGPT exclusively to destroy and refine those drafts from a strict administrative law perspective, the Human Coherence Anchor can extract an output far superior to what any single model could produce alone.

Appendix II: The Consultation Sample

To ground our analysis of the ESRS Delegated Act and Article 49(3b), our Hybrid Intelligent Team (HIT) processed 60 individual comment submissions representing 67 unique organizations. While this represents a convenience sample rather than a comprehensive audit of the 453 total submissions received by the European Commission, it captures a critical cross-section of the most influential stakeholders shaping European corporate governance. As noted in the table below, we have placed these organizations into broad functional groupings that align with the analytical framework of our project.

Appendix III: The Article 49(3b) Vulnerabilities

To unpack the friction between the European Commission’s political mandates and EFRAG’s technical advice, we utilized our Hybrid Intelligent Team (HIT) to sample the public consultation letters on the simplified ESRS Delegated Act. This epistemological triangulation experiment reconstructed a framework of regulatory and institutional vulnerabilities surrounding Article 49(3b).

Our analysis categorizes the potential red flags into six broad areas:

1. Procedural & Due Process Failures: Examining the friction between the Commission’s substantive alterations of EFRAG’s technical advice and the statutory requirements for transparent, evidence-based due process.

2. Scope, Boundaries, and Carve-Outs: Evaluating the legal and evidentiary consequences of exempting specific financial actors or setting rigid size thresholds within the reporting value chain.

3. Measurement, Estimation, and Reliefs: Assessing the tension between decision-useful financial disclosures and the expanding use of conditional omissions, phase-ins, and qualitative reporting reliefs.

4. Assurance & Auditability Constraints: Highlighting the liability traps created for the accounting profession when subjective compliance criteria and unbounded stakeholder obligations are subjected to limited assurance requirements.

5. Interoperability & Global Friction: Analyzing the gap between technical alignment and true legal equivalence with the global ISSB baseline, particularly regarding commercial sensitivity and fragmented financial mandates.

6. Institutional Governance & Judicial Review: Questioning the functional independence of the advisory process, the allocation of authority, and the practical challenges of standing and admissibility for administrative legal challenges.

Because our models engaged in a rigorous Red Team review to verify the epistemological boundaries of this data, we are publishing the full, 3,800-word regulatory analysis as a standalone piece later this week. That forthcoming publication will include a detailed breakdown of the specific provisions and comprehensive tables mapping each of these vulnerabilities to the exact stakeholder groups (Preparers, Auditors, Investors, and Civil Society) that bear the exposure.

Appendix IV: Dramatis Personae (from William Shakespeare’s Hamlet)

Claudius (DG FISMA) The ambitious and calculating King of Denmark who seized the throne by secretly murdering his own brother. He hastily marries his brother’s widow, Queen Gertrude, and uses political maneuvering to maintain his grip on power. Guilt-ridden yet unrepentant, he constantly schemes to eliminate his nephew Hamlet to protect his ill-gotten crown.

Hamlet (EFRAG) The melancholy Prince of Denmark who is tasked by his father’s spirit to avenge his foul and unnatural murder. Deeply philosophical and paralyzed by moral doubt, he feigns madness to uncover the truth of his uncle’s treason. His continuous hesitation and brooding introspection ultimately lead to a tragic cascade of deaths within the royal court.

Polonius (Auditors) The pompous and long-winded Lord Chamberlain who serves as the chief counselor to King Claudius. He relies on espionage, eavesdropping, and political meddling to secure his family’s standing in the Danish court. His habit of hiding behind tapestries to spy on others leads directly to his sudden and violent death.

Laertes (NGOs) The passionate and fiercely loyal son of Polonius who serves as a direct, action-oriented foil to Prince Hamlet. Upon learning of his father’s murder, he immediately raises a rebellion and storms the castle demanding swift vengeance. He readily agrees to use a poisoned sword in a rigged fencing match, though his treachery ultimately claims his own life.

Ophelia (Companies) The dutiful and obedient daughter of Polonius whose tragic romance with Hamlet is destroyed by court politics. Torn between her family’s manipulative commands and the erratic cruelty of her former lover, her mind completely fractures under the strain. She descends into madness, distributing symbolic flowers to the court before drowning in a willow-shaded brook.

Horatio (Investors) A rational, steadfast scholar from Wittenberg who remains Hamlet’s one true and unshakable confidant. He serves as a calm, objective witness to the supernatural and political horrors that consume the royal family. As the sole survivor of the final bloodbath, he is entrusted to remain alive to tell the prince’s tragic story to the world.

The Ghost (The Green Deal) The restless and formidable spirit of the deceased King Hamlet, condemned to walk the night in purgatorial torment. He appears to the castle guards and his son to reveal the shocking truth of his assassination by poison. He lays an uncompromising mandate of revenge upon Prince Hamlet, sparking the entire chain of tragic events.

Fortinbras (IFRS Foundation) The decisive, militaristic Prince of Norway whose own father was slain in combat by the late King Hamlet. He spends the duration of the play marching his army across the background, focused entirely on reclaiming his nation’s lost territories. Arriving just as the Danish royal court destroys itself, he effortlessly inherits the empty throne and restores order to the realm.

Professor Kerstin Lopatta became Chair of EFRAG’s Sustainability Reporting Board on May 1, 2026, for what was supposed to be a three-year term. Five months later, on September 30, 2026, she resigned. Corporate Disclosures reported on October 1, citing Börsen-Zeitung, that disagreements over EFRAG governance—and specifically tensions between the Sustainability Reporting Board and the Sustainability Reporting Technical Expert Group over authority in the standard-setting process—were decisive in her departure.

That account remains reported rather than independently established, and it should be treated as such. Nor is Lopatta a detached witness to the history criticized in this essay. She joined the SRB in 2022, served as Acting Chair during completion of the first set of draft ESRS, and then spent years as Vice-Chair before becoming Chair. That makes her position more complicated, not less relevant. The governance question is why someone who helped carry the original process through, remained inside the institution and later became the accountable Board chair left after five months amid a reported dispute over authority.

Overview

The resignation of Lopatta on September 30—and the accompanying reports of deep governance failures and unapproved board records—has fractured the consensus surrounding Europe’s sustainability reporting regime. What is frequently dismissed in Brussels as a dispute over technical standard-setting has instead escalated into a structural crisis for the European Commission and the operability of its flagship Corporate Sustainability Reporting Directive (CSRD).

This essay is admittedly long and deeply technical, but that complexity is necessary to unpack a regulatory architecture that was hardwired to prevent legislative correction. Because the mechanics of that architecture matter, the primary arguments can be summarized as follows:

First, Lopatta’s departure is not a localized personality conflict. It exposes a fundamental breakdown in the corporate governance of EFRAG, an institution whose technical apparatus has consistently bypassed or overpowered its own board oversight.

Second, this internal governance failure carries immediate legal consequences. Article 49(3b) of the Accounting Directive does not merely suggest good behavior. It makes EFRAG’s “proper due process, public oversight and transparency” a strict statutory condition of the advice the Commission relies upon. Procedural failures at EFRAG therefore render the resulting delegated acts legally vulnerable to challenge by member states.

Finally, the crisis shifts the ultimate burden of accountability directly to the European Commission. The Directorate-General for Financial Stability, Financial Services and Capital Markets Union (DG FISMA) provides roughly 90 percent of EFRAG’s funding, yet it has failed to exercise institutional oversight commensurate with the systemic, cross-border regulatory risks that EFRAG’s standard-setting architecture created.

The Bigger Picture

The stakes are larger than one resignation or even EFRAG itself. Mario Draghi’s competitiveness report put regulatory burden squarely inside Europe’s growth problem, and an economically stagnant Europe whose corporate champions divert managerial attention and capital from innovation to administrative compliance is an acute strategic liability for the democratic alliance. The capital market numbers tell the real story. The European Commission itself notes that EU stock-market capitalization was only 73 percent of GDP in 2024, compared with 270 percent in the United States. Across more than 300 fragmented exchanges, the combined capitalization of the entire European Union sits around $12 trillion—roughly one-fifth the market capitalization of the two largest U.S. exchanges alone.

Europe is now trying to reconcile two policy eras: one that built an extraordinarily ambitious sustainability-reporting architecture to support the European Green Deal and its goal of achieving net-zero emissions by 2050, and another that is explicitly charged with simplification, productivity, and reducing reporting burdens. Valdis Dombrovskis now serves as the European Commissioner for Economy and Productivity, Implementation and Simplification, working alongside the Directorate-General for Economic and Financial Affairs (DG ECFIN). Maria Luís Albuquerque serves as the European Commissioner for Financial Services and the Savings and Investments Union, overseeing the Directorate-General for Financial Stability, Financial Services and Capital Markets Union (DG FISMA), whose portfolio includes the sustainability-reporting file. She has in fairness demanded simplification, usability, and proportionality from EFRAG. The institutional tension is therefore not a morality play between reformers and obstructionists. It is a test of whether a regulatory system built under one set of assumptions can be governed credibly when the political and economic assumptions change.

In my July 2022 piece “The Credibility of EFRAG’s Sustainability Reporting Standards Is At Risk” I raised concerns about EFRAG’s process where participants described a “hot mess” of compressed timetables, obscured disagreement, and people being steamrollered. Last week in “The Missing Whistleblower, the Ghosted Chair, and Wired Mice” I raised the governance issue again and asked the central question of whether this institution has governance machinery strong enough to surface and adjudicate disagreement before a resignation becomes the mechanism of disclosure. Four years later, that is no longer a retrospective complaint about the difficult birth of European Sustainability Reporting Standards (ESRS). It is a present-tense institutional question: whether ESRS can survive EFRAG.

EFRAG and ESRS: Key Institutional Facts

The story is a complicated one, so Table 1 summarizes the key issues I’ll be discussing.

How EFRAG Got the Pen

The policy problem that produced the Corporate Sustainability Reporting Directive (CSRD), passed in November 2022 and in force starting in January 2023, was real. The Non-Financial Reporting Directive had already introduced the two-sided logic later called double materiality, but its voluntary guidance did not produce information of sufficient comparability or consistency. The CSRD therefore moved from broad disclosure principles toward standardized European Sustainability Reporting Standards, with EFRAG designated to prepare the technical advice on which the Commission would base delegated acts.

There was a defensible theory behind that move. Markets cannot price information they do not possess; environmental and social impacts can be economically consequential; standardized information can reduce search costs and expose greenwashing. The mistake was to let disclosure carry more policy weight than disclosure can bear. Disclosure is an informational technology. It is not an energy technology. It cannot substitute for competitive power prices, permitting reform, deep capital markets, scalable venture finance, transmission infrastructure or an industrial environment in which firms can earn adequate returns on green investment.

That distinction matters because standard-setting is not the same activity as policy advocacy although many NGOs confuse the two. A public-interest constituency may reasonably ask for more information about corporate impacts. A standard setter must ask a harder set of questions before turning that demand into an auditable requirement: Can the information be defined and measured consistently? Can management establish controls over it? Can an assurance provider obtain evidence? Will sophisticated users actually use it? What does it cost? How does it behave across sectors and jurisdictions? What happens when the information lies three tiers down a value chain, when foreign law constrains collection, or when the company cannot reasonably obtain it? Those questions do not weaken sustainability reporting. They determine whether it functions.

EFRAG was not originally built for this role. It was established in 2001 as an association internationale sans but lucratif (AISBL)—an international non-profit association governed by Belgian law—whose core task was to advise the European Commission on whether standards issued by the International Accounting Standards Board (IASB), established on April 1, 2001, should be endorsed for use in the Union. Its role was derivative and reactive. Another institution wrote the standards; EFRAG evaluated them. The 2004 IAS 39 controversy later confirmed and intensified a European debate over accounting sovereignty, when EU endorsement included carve-outs after sustained objections, particularly from France to elements of the IASB’s fair-value approach. The episode entrenched a legitimate concern that endorsement of standards written elsewhere was not the same thing as having institutional control over their design. Sustainability reporting created a field in which Europe could be an author rather than an endorser. As Mairead McGuinness—Maria Luís Albuquerque’s predecessor as Commissioner for Financial Services—summarized the political ambition during the Parliament’s debate on the directive, Europe intended to be not a “rule-taker but a rule-maker.”

The critical institutional move came in June 2020. Then-Executive Vice-President of the European Commission for An Economy that Works for People Valdis Dombrovskis gave Jean-Paul Gauzès—a former French Member of the European Parliament and then-president of EFRAG—an ad personam mandate to recommend the governance changes EFRAG would require if it were entrusted with sustainability standard-setting. In parallel, the Commission asked EFRAG to undertake preparatory technical work on possible EU non-financial reporting standards. EFRAG carried that work out through a Project Task Force chaired by Patrick de Cambourg, president of the French accounting standard setter Autorité des normes comptables (ANC). Both exercises reported in March 2021. The problem is not that the Commission literally failed to consider alternatives; its consultation and impact-assessment process did address standardization options and institutional arrangements.

The sharper point is that the options analysis ran in parallel with EFRAG’s own president designing EFRAG’s future governance and EFRAG’s own task force designing the future standards. By the time institutional choice was being tested, the favored institution was already designing the job and doing the work. EFRAG was subsequently written directly into the draft legislation. Because votes on European directives in the Parliament and the Council are ultimately up-or-down decisions on an entire negotiated compromise package, EFRAG’s role was wired into the law from the start, leaving lawmakers no practical way to challenge the standard-setter without sinking the directive itself.

The Commission’s CSRD proposal followed on April 21, 2021, and the provisional trilogue (the closed-door negotiations where the Commission, Parliament, and Council finalize EU legislation) agreement on June 21, 2022. The co-legislators did not simply give EFRAG an unchecked mandate. They wrote safeguards into what became Article 49(3b) of the Accounting Directive. In its current form, the Commission is to take EFRAG’s technical advice into consideration only where, among other things, it was developed with “proper due process, public oversight and transparency,” with balanced expertise and participation, with “sufficient public funding to ensure its independence,” and on a “work programme on which the Commission has been consulted.” The advice must also be accompanied by a cost-benefit analysis. Those conditions survived Omnibus I—the legislative package introduced to simplify corporate obligations and narrow reporting thresholds—and remain fully intact in the consolidated Accounting Directive, which integrates the original 2013 statute and all subsequent amendments, including the CSRD, into a single binding text.

This is not a nuisance qualification to the governance argument. It is its legal core. The legislature made the quality and independence of the process a statutory condition attached to the advice on which the Commission relies. EFRAG’s first exposure drafts show why that safeguard matters. While consultation and Commission intervention ultimately corrected early overreach—such as the withdrawn rebuttable presumption of materiality—a defender of the existing system might argue it operated as designed: EFRAG advised, others commented, and the Commission decided. But the sheer volume of correction, simplification, and re-engineering required asks a fundamental question:. Why was so much fixing necessary after the advice had already passed through a body whose statutory relevance depends on due process and public oversight? The issue is not that EFRAG possesses final legal authority. It is whether it has exercised its technical authority with governance commensurate to the consequences.

The Record of Reversal

The chronology is revealing because it shows repeated contraction after ambitious initial design. EFRAG has demonstrated that it can cut when mandated to do so. The problem is that these cuts consistently arrive in response to outside pressure—consultation backlash, Commission intervention, or explicit legislative mandates—rather than from a self-initiated conclusion that the architecture had become too heavy. Table 2 below acts as an evidentiary record of internal governance failing to self-correct until politicians or co-legislators step in to interrupt the implementation timetable.

Revision is not itself evidence of failure. Serious standards evolve. But this degree of instability has a severe cost, especially when companies have already built systems, hired advisers, configured software, redesigned controls, and imposed data requests down their supply chains. Repeated reconstruction acts as a levy on preparers. It destroys the predictability that standard-setters are supposed to provide, eroding confidence that today’s expensive compliance build will still be required tomorrow. The fact that EFRAG ultimately delivered major simplification, albeit after four years of work and pressure, when explicitly told to do so is welcome. It is also the measure of how much burden had accumulated before the political system forced the recalibration.

What the Pen Produced: E1-11 and Article 40a

Two current design problems illustrate why governance cannot be separated from technical quality. Evaluating this governance, however, requires looking at two distinct levels: the internal governance of EFRAG itself, and the external oversight exercised by DG FISMA, which provides 90 percent of EFRAG’s funding. Neither design problem should be caricatured as proof that EFRAG alone caused the result, as EFRAG is constrained by Level 1 legislation—the foundational statutory directives passed by the European Parliament and Council—and the Commission retains the authority to alter EFRAG’s advice. The question is whether a properly governed technical process identifies operability problems early, records dissent, tests alternatives in real companies, and makes the consequences unmistakable to the institution that will turn advice into law.

Anticipated Financial Effects

The first example is anticipated financial effects (AFE) in ESRS E1-9. The 2023 standards already combined time-based phase-ins with an entity-specific test. ESRS E1-9 could be omitted entirely in the first year, and for the first three years an undertaking could provide only qualitative information if quantitative disclosure was impracticable. The revised 2026 architecture should therefore not be described as a move from a calendar test to a capability test. It replaced one demanding entity-specific test with a different, permanent one while retaining a separate phase-in. Revised ESRS 2 paragraph 29 says an undertaking need not provide quantitative anticipated-financial effects information if it lacks the skills, capabilities, or resources to do so. That relief does not expire in 2030. What ends then, for the relevant wave-one undertakings (large public-interest entities already subject to the NFRD), is the broader transitional permission to omit specified quantitative information. After 2030, the pressure therefore comes from application of the standing capability test—not from the disappearance of the test itself.

Borrowing that language from IFRS S2 is not inherently objectionable; interoperability argues for it. The harder question is how the same clause operates in a population concentrated, after Omnibus I, on very large undertakings. In IFRS S2 the proportionality mechanism must work across issuers of very different sizes and levels of reporting maturity. In revised ESRS it is available permanently to large companies as well, but its practical force depends on whether a company can substantiate that it still lacks the relevant skills, capabilities, or resources. That is a judgment auditors and supervisors may test, but it is not a calendar rule and should not be presented as one. Nor is the fact that estimates may later change an indictment of the standard. Revised ESRS expressly treats estimate revisions arising as additional information becomes known as distinct from prior-period errors, consistent with the general estimates-versus-errors logic already found in ESRS 1 and IAS 8. The serious operability question is narrower: whether the required E1-11 quantification can be produced with sufficient decision-usefulness, control and assurance across the large-company population, and whether the permanent relief is applied consistently enough to avoid arbitrary differences between similarly situated preparers.

The technocratic framing obscures why AFE represents such a profound governance failure. The irony is staggering. Under an Omnibus mandate explicitly designed to deliver simplification, EFRAG quietly introduced a massive increase in modeling complexity. In the original standards, AFE (then located in E1-9) was a broader requirement covering multiple sustainability topics, including water and the circular economy. In the revised standards, AFE was shifted to E1-11 and narrowed specifically to climate but made exponentially more demanding. It now requires highly detailed quantitative modeling of financial effects on an undertaking’s assets and revenues under a strict 1.5°C scenario.

This brings us to the core due process question. EFRAG essentially snuck this complex, materially different 1.5°C financial modeling architecture into the final text without re-exposing it for public comment. The relevant standard here is not an American “logical outgrowth” doctrine transplanted into EU law; it is EFRAG’s own due-process framework read together with the statutory safeguards of Article 49(3b). Those safeguards obligate the European Commission’s DG FISMA to actively police standard-setting hygiene, not merely rubber-stamp the output. If a standard-setter uses a simplification mandate to quietly hardwire a vastly more complex requirement into the rules, the record must show who authorized it and what testing supported it. Formal adoption by the Commission does not cure the fact that EFRAG bypassed transparency; rather, it exposes a critical breakdown in DG FISMA’s oversight, raising the uncomfortable question of whether the Commission even understood the technical payload it was adopting. Respondents were denied a meaningful opportunity to challenge this new architecture before it was cemented into law. Table 3 below compares how AFE was handled in the two sets of standards.

Article 40a

Article 40a raises a different operability problem. As revised by Omnibus I, the extraterritorial mandate applies to non-EU parent groups that generate more than €450 million in EU turnover and have at least one EU subsidiary or branch generating more than €200 million. For these qualifying groups, the law requires a sustainability report at the ultimate-parent level under dedicated standards. But Article 40a does not simply apply “impact materiality” as a named test. Its statutory content is drawn from the impact-related side of the CSRD architecture: the third-country report is confined to impacts rather than the full package of impacts, risks, and opportunities. That means reliefs designed around anticipated financial effects have no object in the Article 40a report.

The important asymmetry is therefore operational. A European subsidiary within full ESRS applies double materiality. Its third-country parent under Article 40a is being asked to report on impacts whose perimeter is defined by business relationships and causal attribution rather than financial reporting control. That perimeter is not literally unbounded: severity and likelihood govern impact materiality, value chain information is conditioned by reasonable and supportable information, and the Omnibus reforms seek to limit trickle-down demands on smaller partners. But the remaining burden can still be substantial. Consider workers in the value chain. A multinational may need to identify material labor impacts, describe policies and engagement, explain remediation and actions, and support targets across upstream or downstream relationships that do not sit inside the accounting consolidation perimeter. That is a real information-system problem even where the legal design is defended as proportionate to access to the EU market.

The July 2026 ESRS-40a exposure draft proposes a mixed approach rather than settling one. Under the proposal, reporting is generally global, while a group may elect to limit non-climate disclosure to EU-related impacts; climate remains global. That distinction should be debated in the consultation, which remains open through October 31, not described as enacted law. It also changes the interoperability analysis. IFRS S2 already requires Scope 1, 2 and 3 greenhouse-gas emissions across the value chain, so climate inventories and related systems are substantially reusable. For non-climate impacts, however, the more relevant comparator is the Global Reporting Initiative (GRI), not the ISSB, because the ISSB has not issued topic standards covering the social and other non-climate impact disclosures at issue.

Many large U.S. multinationals already report with reference to, or in accordance with GRI, and EFRAG and GRI have rightly emphasized the high level of interoperability between their impact-reporting architectures. That means a serious critique cannot pretend those companies start from zero. But interoperability is not identity. The joint GRI-ESRS mapping itself records differences in scope, granularity, data type, and definitions, while Article 40a would impose a legally defined reporting perimeter and a mandatory European standard rather than a company’s existing voluntary selection of GRI disclosures. The practical issue is therefore conversion, boundary, and completeness—not wholesale duplication. A mature GRI reporter may reuse substantial processes and data, but it may still have to determine which global or EU-related impacts fall within the Article 40a perimeter, close disclosure gaps, document materiality under the European rules, and support a report capable of satisfying the legal and assurance framework. The equivalence route remains relevant, as does the possibility of parent-level full-ESRS reporting that can simplify subsidiary reporting, but neither should be treated as proof that fragmentation disappears.

Interaction Effects

The interaction between Article 40a and the E1-11 safe harbor exposes a profound structural contradiction in the EFRAG architecture. Article 40a forces non-EU multinationals to report exclusively on the impact side of the double-materiality ledger at a global level. The irony here is acute. EFRAG provides a capability-based safe harbor for EU companies modeling financial effects yet demands that third-country multinationals map their global impacts with absolutely no such relief. This is a severe asymmetry, because impact measures are an order of magnitude more difficult to execute. Unlike financial materiality, which relies heavily on internal enterprise data and standard accounting parameters, impact materiality requires sourcing data from far outside the corporate boundaries. It raises profound boundary questions—challenges that the Global Reporting Initiative (GRI) grapples with constantly—and demands deep localized contextualization. For example, the impact of consuming 10,000 liters of water in a severely water-stressed region is vastly different from consuming the same amount where water is plentiful. The standard setter’s architecture effectively demands total, unmitigated global competence on the most difficult part of the equation while feigning leniency on localized financial effects.

When you view this through the lens of a corporate group, the contradiction becomes an interoperability trap. If a U.S. parent company is marshaling the capital and data architecture to produce a comprehensive global impact report under Article 40a, its EU subsidiary can hardly turn to an auditor and credibly invoke the E1-11 safe harbor by claiming it lacks the resources to model local financial climate risks. While possessing global impact capabilities does not automatically guarantee localized financial modeling skills, an assurance provider will inevitably look up the corporate ladder. If the ultimate parent possesses the infrastructure to execute a complex global impact assessment, the assurers will be hard-pressed to accept that the subsidiary is devoid of resources. By forcing the parent to build a massive global reporting apparatus, the Article 40a mandate effectively neutralizes the E1-11 safe harbor for the subsidiary—a structural conflict detailed in Table 4.

The Governance Crisis

EFRAG’s sustainability pillar has a divided constitution. The Sustainability Reporting Board (SRB) carries institutional responsibility for EFRAG’s sustainability-reporting positions and the technical advice submitted to the Commission. The Sustainability Reporting Technical Expert Group (TEG) performs much of the technical work and advises the Board. The constitutional premise is therefore clear enough on paper; the Board decides and the TEG advises. The governance question is whether that division operates in practice when technical authority, deadlines, and institutional power collide.

That is why Lopatta’s departure matters, notwithstanding her own role in the system’s history. A person who helped steer the first standards, remained in the governance structure, and ultimately became the formally accountable SRB Chair resigned after five months amid reported disagreement about the allocation of authority between Board and TEG. In striking contrast, the Chair of the Sustainability Reporting TEG, Chiara Del Prete—whose background includes years as an auditor at KPMG and PwC before becoming a partner at Mazars—recently had her term extended through January 2028. This extension pushes her continuous service as an EFRAG TEG chair to nearly nine years, bypassing the organization’s standard six-year term limit. That juxtaposition does not prove a structural indictment. It does create a serious present-tense question that should be answered through minutes, decision records, and documented authority rather than through either side’s characterization.

The test of governance is not whether an organization has statutes, committees, and due process manuals. It is what happens when expert judgment conflicts with political urgency; when a board disagrees with the technical apparatus that serves it; when a deadline makes genuine reconsideration inconvenient; when dissent is regarded as obstruction; and when the institution must choose between admitting that a concept is not operationally ready and meeting a Commission timetable. If the Board can meaningfully reject, amend or remand TEG work, the minutes and decision record should show it. If it cannot, then the formal allocation of authority is misleading. If the relationship broke down during the simplification process, the breakdown matters because EFRAG was being asked to redesign standards whose legitimacy depends on precisely this governance chain.

The Administrative Board succession should be approached with the same discipline. Benoit Jaspar—a former KPMG manager, executive adviser at Assicurazioni Generali, and lecturer at the Université Libre de Bruxelles—is retiring as Administrative Board President at the end of the year. The issue is not which individual should replace him, nor should an essay like this imply that a particular candidate represents a predetermined outcome. But the candidates themselves are making the structural crisis explicit. Georg Lanfermann, President of the German standard setter DRSC, recently announced his candidacy to replace Jaspar with a public platform explicitly targeting the “growing external cristiscm about EFRAG’s governance.” His campaign presentation formally calls for establishing “clear reporting lines.” reworking the appointment process for Technical Board Chairs,” extending due process oversight to capture “governance deficiencies,” and fostering an open discussion culture “founded on mutual respect.”

When a leading candidate for the presidency builds his platform on the need to fix broken reporting lines and restore basic institutional respect, a transparent succession process is not a side issue. It is a test of whether EFRAG understands that legitimacy depends on demonstrable procedure, not confidence that insiders acted appropriately.

None of this requires assuming bad faith. Institutions can develop blurred lines of authority without conspiracies; technical groups can become dominant because they possess information, continuity, and staff support; boards can ratify work they formally control because deadlines make reopening it costly. Those are familiar organizational dynamics. They are also exactly why independent governance review exists. The question is not whether EFRAG’s organizational chart is defensible. It is whether authority, dissent, and accountability operate in practice as the chart says they do.

If Brussels Does Not Act

Brussels should consider what this record looks like to adversaries rather than colleagues. When corporate reporting rules move from voluntary guidance to hard law carrying severe liability, the theater of dispute inevitably shifts from standard-setting working groups to the courts. If a coalition of multinational corporations and transatlantic trade associations decides that the costs, ambiguities, and extraterritorial overreach of the revised ESRS are no longer acceptable, it will not engage through polite public feedback periods. It will deploy top-flight European public-law counsel and regulatory litigators.

The primary target in their crosshairs will not be broad political grievances, but the statutory trap embedded in Article 49(3b) of the Accounting Directive.

The Legal Artillery: Weaponizing Article 49(3b)

Article 49(3b) is not an aspirational preamble. It establishes mandatory statutory preconditions that the European Commission must satisfy before adopting delegated acts. The technical advice it relies upon must be developed under “proper due process, public oversight and transparency,” with balanced stakeholder representation, robust cost-benefit analysis, and demonstrably independent technical work. While the Commission inserted boilerplate recitals into Delegated Regulation (EU) 2026/1563 asserting that EFRAG complied with these criteria, a regulatory recital is merely an assertion—it is not an adjudicated fact immune from judicial scrutiny.

A serious legal assault would not waste time on fragile direct annulment claims before the General Court under Article 263 TFEU (Treaty on the Functioning of the European Union), where standing hurdles for private litigants are notoriously steep. Instead, the devastating route runs through the national courts via an Article 267 preliminary reference to the Court of Justice of the European Union (CJEU):

The Procedural Vehicle: A multinational or affected trade body challenges a national enforcement notice, transposition mandate, or audit requirement in a member state court (such as Germany, France, or the Netherlands), alleging that the underlying delegated act is void because the Commission acted ultra vires by relying on advice produced in direct violation of Article 49(3b).

The Evidentiary Pipeline: European administrative litigation does not have broad, American-style discovery or depositions, but challengers have equally lethal mechanisms to expose institutional dysfunction. Coordinated, massive access-to-documents requests under Regulation (EC) 1049/2001 targeting DG FISMA and Commission correspondence, paired with formal complaints to the European Ombudsman for maladministration, would force internal records into the light. Unapproved board minutes, documented procedural overrides of the Sustainability Reporting Board by the TEG, and internal dissents regarding un-consulted additions (like the 1.5°C anticipated financial effects mandate) would become direct evidentiary exhibits before the Court.

Timeline and Market Paralysis: An Article 267 preliminary reference is not a quick skirmish; it is an 18-to-24-month judicial siege that can easily stretch into three or four years before a final ruling from Luxembourg. (Article 267 establishes the “preliminary reference” procedure, allowing national courts in member states to pause domestic proceedings and ask the Court of Justice of the European Union to definitively interpret EU law or rule on the validity of EU regulatory acts before deciding a domestic case.) While obtaining formal interim relief to suspend an EU regulation under Article 278 TFEU requires meeting exceptionally high legal bars, the practical consequence is immediate paralysis. Boards, general counsels, and audit committees confronted with active, high-stakes litigation challenging the very validity of the ESRS will hesitate to commit tens of millions of euros into building systems to comply with standards that may be struck down.

The Internal Revolt: The 17-State Transposition Shield

The litigation threat does not exist in a vacuum; it lands in an EU where domestic political support for the CSRD has fractured. More than half the bloc—17 out of the EU’s 27 member states—failed to transpose the original directive by the July 2024 deadline and face formal European Commission infringement proceedings. The holdout list is striking. It includes Portugal, providing the severe political irony that the home country of Maria Luís Albuquerque has itself stalled on implementation. It also includes major economic players like Germany, the Netherlands, and Belgium (home of EFRAG). Having missed the initial two-month window to respond to formal notices, these nations trigger the Commission’s authority to issue “reasoned opinions,” the final warning before referral to the European Court of Justice. If referred, Article 260(3) TFEU empowers the Court to levy severe daily financial penalties and lump sums. For an economic heavyweight like Germany, the Commission’s penalty calculation multiplier—the ‘N’ factor, a metric based on a country’s gross domestic product and number of European Parliament seats—is the highest in the bloc at 6.02, ensuring any daily fines demanded by Brussels would be uniquely punishing.

But the Commission’s enforcement weapon can cut both ways. If Brussels attempts to financially coerce transposition, holdout nations possess a structural countermeasure: suing the Commission first. Under Article 263 TFEU, Member States are classified as “privileged applicants,” granting them automatic legal standing to seek the annulment of EU acts without having to prove individual harm. Rather than passively absorbing daily infringement fines, a country like Germany could launch an action for annulment against the Commission’s delegated act, leveraging the exact due-process failures EFRAG committed. The legal grounds for an Article 263 challenge map perfectly onto EFRAG’s bypassed transparency: procedural irregularity, lack of competence, and the infringement of statutory safeguards like those outlined in Article 49(3b). If Brussels attempts to force transposition through financial penalties, an aggressive Member State might simply force the European Court of Justice to adjudicate how a materially different 1.5°C architecture was smuggled into the final law.

The 40a Interoperability Collision

This internal member state defiance collides directly with the external pressure mounting across the Atlantic. Under the EU–U.S. trade framework, Washington has placed the extraterritorial reach of the CSRD squarely on the negotiating table. If Brussels bows to this bilateral pressure and offers third-country concessions—such as carving out non-EU groups from Article 40a or downgrading their global impact reporting to a voluntary status—European industry will erupt.

Domestic European companies, already bearing the full weight of double materiality, will not tolerate a regime where U.S. competitors operating within the single market are granted exemptions while local firms remain chained to burdensome disclosure mandates. Cries of “reverse discrimination” and and unlevel playing fields will render the current framework politically unviable inside Europe.

Caught between the threat of protracted judicial review under Article 49(3b), open rebellion from 17 member states refusing to implement the regime, and the commercial impossibility of offering one-sided reliefs to foreign multinationals, the Commission’s defensive perimeter will collapse. When an administrative architecture becomes this brittle, half-measures and minor technical amendments will no longer suffice. Brussels will be forced to reckon with the reality that the only path out of this institutional gridlock is a fundamental retreat: establishing a workable, interoperable global baseline anchored firmly in financial materiality, while fundamentally reconsidering whether the expansive, highly contested architecture of impact materiality can ever function as a mandatory statutory regime.

Figure 1 summarizes the forces at work.

What the Commission Should Do

The answer is not an EFRAG self-assessment, a lessons-learned exercise whose terms of reference are negotiated by the institution being examined, a stakeholder roundtable, or a review confined to whether written procedures were formally observed. Something much stronger than this is needed. The appropriate response is a genuinely independent governance review commissioned at a level capable of examining both EFRAG and the Commission’s oversight of it.

Europe has a direct precedent. In March 2013, Michel Barnier, then-European Commissioner for Internal Market and Services, appointed Philippe Maystadt—former Belgian Deputy Prime Minister, Minister of Finance, and President of the European Investment Bank—as a Special Adviser to examine EFRAG’s governance and recommend structural reforms to reinforce the EU’s role in international financial reporting. Maystadt’s report led directly to substantial institutional reform. The structural distinction matters. Maystadt was appointed by the Commission as an external special adviser; Jean-Paul Gauzès received the 2020 governance mandate while serving as EFRAG’s own President. An external review in 2026 would therefore not be an attack on EFRAG. It would be an established European method of protecting the legitimacy of an organization whose role has materially changed.

The review should examine the allocation of authority among the Sustainability Reporting Board, the Sustainability Reporting Technical Expert Group, the Secretariat, and the Administrative Board; the operation of the Due Process Committee; how material dissent is recorded and escalated; whether boards can in practice amend, reject, or remand technical work; the handling of major changes between exposure drafts, technical advice, and the final Commission text; the design and use of field testing and cost-benefit analysis; and the Commission’s own internal process for testing whether Article 49(3b)’s statutory conditions have been satisfied.

Crucially, this review must audit not just organizational charts, but how executive power and personnel are managed at the apex of the institution. The current ambiguity regarding leadership selection, compensation, and oversight requires a formal review framework (Table 5).

The independent review should also reconstruct several consequential standard-setting decisions from beginning to end—not to retry every technical judgment, but to establish who decided what, under which authority, and after what evidence and consultation. The complete report should be formally submitted to the European Parliament’s Committee on Economic and Monetary Affairs (ECON) and the Council, with the final findings published directly on DG FISMA’s official portal and in the European Union’s public document register, with legitimate commercial or personal confidentiality protected narrowly rather than used to obscure institutional facts.

DG FISMA’s oversight belongs inside the review because a standard setter does not operate in an institutional vacuum when its advice becomes the basis for delegated legislation. The Commission must be able to show that it has not outsourced politically consequential normative choices and then treated technical advice as presumptively legitimate because EFRAG complied with procedures EFRAG itself administers.

EFRAG also receives substantial Union financial support. That makes the European Court of Auditors—the European Union’s independent external auditor tasked with safeguarding sound financial management and scrutinizing the use of EU funds—a relevant institutional backstop on questions of EU financing and Commission oversight of grant recipients, even though the Court is not a technical appeals body for ESRS. The European Ombudsman—the independent body established to investigate maladministration, procedural unfairness, and lack of transparency across EU institutions and agencies—could likewise examine whether DG FISMA’s administrative oversight met required standards of diligence when complaints are brought before it. None of these routes substitutes for an independent review; their relevance is a reason for the Commission to act before external scrutiny becomes the mechanism of reform.

The historical responsibility is unusually direct. Valdis Dombrovskis helped initiate the governance architecture in 2020 when Europe wanted a sustainability standard setter. He now holds the Commission portfolio charged with implementation, simplification, reporting burden reduction, elimination of regulatory contradiction, and better law-making. He does not need to repudiate the CSRD, reject double materiality, or take sides in every internal EFRAG disagreement. He needs to ask a more basic question. Does an organization exercising this degree of influence over European corporate regulation have decision-making structures commensurate with that responsibility?

That is not an argument for weaker sustainability reporting. It is an argument for reporting strong enough to survive contact with preparers, auditors, investors, courts, trade officials, and the political economy in which the standards must operate. The IFRS Foundation—and its dual-board structure comprising the International Accounting Standards Board (IASB) and the International Sustainability Standards Board (ISSB)—offers an immediate comparative baseline. EFRAG does not need to copy the IFRS Foundation’s governance manuals verbatim, but it must reckon with its institutional credibility. The ISSB has successfully established a global baseline for financial materiality largely because its due process, board oversight, and executive boundaries are structurally legible to the markets it regulates. Europe’s ambition to lead on impact materiality cannot rely on a governance architecture that is less mature, less transparent, and less accountable than the global baseline it seeks to surpass.

EFRAG helped build ESRS. An independent review would protect both the institution and the standards from the consequences of unresolved doubt about authority, process, and operability. Dombrovskis now occupies precisely the office capable of asking whether the structure he helped set in motion still deserves Europe’s confidence. He should use it.

The last five years have witnessed enormous anti-ESG pushback from elected Republican officials in red states. Examples include a letter to asset managers from 25 red state attorneys general and treasurers, which my Republican buddy Dan Crowley and I wrote about here, and another from 16 red state AGs attacking the Big Four accounting firms, which Dan and I wrote about here. Because these missives are primarily performative political theater rather than economic substance, they usually make for a highly entertaining read. Channeling Macbeth, they’re the old “full of sound and fury, signifying nothing” kind of thing.

In the meantime, companies press on with their work on sustainability and sustainability reporting. According to the Governance & Accountability Institute’s latest 2026 research, 94% of the S&P 500 and 86% of the Russell 1000 published sustainability reports last year, with many voluntarily aligning with the Global Reporting Initiative (GRI). Today, as all of you know, a new architecture of sustainability reporting standards is actively being set. For decades we have had unified standards for financial reporting, but not for sustainability. Their time has come! One of the central organizations setting those standards today is EFRAG—The Artist formerly known as the European Financial Reporting Advisory Group. It was founded in 2001.

Through the Commission Delegated Regulation (EU) 2023/2772, enacted under the Corporate Sustainability Reporting Directive (CSRD), the European Commission formally delegated to this Belgian NGO—technically an Association Internationale Sans But Lucratif (AISBL)—the responsibility for drafting the European Sustainability Reporting Standards (ESRS). EU companies exceeding certain thresholds (generally 1,000 employees and €50 million in turnover, both public and private) are captured. Furthermore, under the recently updated Article 40a thresholds, non-EU foreign parent companies generating over €450 million in the EU with a subsidiary or branch generating over €200 million will also be required to report according to these standards. In total, the CSRD is estimated to capture roughly 10,000 public and private companies globally.

One distinctive feature of ESRS, compared to the framework developed by the International Sustainability Standards Board (ISSB), is that ESRS is based on “double materiality.” This means companies must report on financially material sustainability issues that matter to investors (the sole focus of the ISSB), but also on impact materiality, i.e., the positive and negative externalities the company is creating in the world. This is a highly contentious topic, especially in the U.S. where many Republicans are still struggling with the simple concept of financial materiality (a concept ISSB Vice-Chair Sue Lloyd and I discussed in a recent interview).

That said, like any political regime, the EU is free to pass whatever laws it wants. But the credibility of these laws ultimately depends upon political support and the legitimacy of the tools created for their implementation.

Thus, it was with palpable anticipation that I sat down to read EFRAG’s 2025 Sustainability Report, released just a few months ago. It is their first one! I was eager to see how this organization was putting into practice the rigorous standards it is setting for thousands of companies in Europe and around the world. In colloquial American terms, I wanted to see how they were “eating their own dog food.”

I was left in breathless wonder by the end of its 31 pages. If I were a conspiracy theorist, I would say EFRAG is a Manchurian Candidate secretly funded by the anti-ESG forces in America to discredit European standard-setting from the inside. It is with some rueful regret that I even mutter this speculation, but EFRAG seems to fall squarely in the “Do as I say, not as I do” camp.

Since we are talking about Europe, I want to start by being Continentally gracious. I will not fault an organization of 59 people for choosing to shield itself by reporting under the much lighter Voluntary Sustainability Reporting Standard for SMEs (VSME), rather than applying the bruising, full-scale ESRS requirements (which started with around 2,500 data points, now whittled down to a still healthy 350) that the law will soon require of larger companies. But digging into the substance of what they did report, here are the three takeaways that left me shaking my head.

No Code of Conduct and No Whistleblower Function

Under ESRS, EFRAG demands that companies rigorously disclose the mechanisms by which their workforce can raise concerns (ESRS S1-3), their structures for social dialogue (ESRS S1-8), and the granular details of their corporate culture and whistleblowing channels (ESRS G1-1).

How does EFRAG measure up to its own demands? In disclosure C06.000, which covers human rights policies and processes, EFRAG openly states:

• “Undertaking has a code of conduct or human rights policy for its own workforce: NO”

• “Undertaking has a complaint handling mechanism for its own workforce: NO”

This is not just breathtaking institutional hypocrisy; it is a blatant violation of host-country law. In disclosure B01.000, EFRAG officially reports a headcount of 59 employees (with 54 based in Belgium), alongside an additional 22 self-employed contractors. (You will be pleased to know that all employees are paid above the minimum wage.) Under the Belgian Loi du 28 novembre 2022 (which transposed EU Directive 2019/1937), any private sector legal entity with 50 or more workers is legally mandated to establish a secure, confidential internal reporting channel for whistleblowers. Furthermore, Belgian labor law requires enterprises with 50 or more employees to establish a Committee for Prevention and Protection at Work (CPPT) to oversee employee well-being and social dialogue. EFRAG has blown straight past the statutory threshold and openly admitted they have no mechanisms in place. It’s not a choice; it’s the law. And they are violating it. In fairness, they say they’re going to put this stuff in place this year. Thats like saying, “Come on, man, give me a break. I’ll follow the law when I get around to it.”

The Ghosting of the Sustainability Reporting Board Chair

Perhaps the strangest mystery in the report is found on page 15, which explains in about a sixth of a page EFRAG’s governance structure. When identifying the “Most senior level within its employees that is accountable for policies,” the report states:

“The EFRAG Administrative Board holds the responsibility for policy adoption. The EFRAG CEO and EFRAG SR TEG and FR TEG Chairs hold the responsibility for policy implementation.”

The Chairs of the Sustainability Reporting Board (SRB) and the Financial Reporting Board (FRB) are completely ghosted. The board Chairs are nominated through a rigorous, high-profile process involving the European Parliament and the Commission. It is the SRB Chair who has the ultimate accountability for the standards produced by EFRAG. Furthermore, Article 49(3b) of the Accounting Directive states that the Commission can only adopt these standards if they have been developed with proper due process, public oversight, and transparency.

Yet, in their own operational reality, the governance structure explicitly bypasses the SRB Chair, routing policy implementation entirely through the CEO and the Technical Expert Group (TEG) Chairs. It is as if the leadership role responsible for the sustainability standards does not actually exist within EFRAG’s internal machinery. Perhaps EFRAG’s Administrative Board is feeling a little bit skittish and didn’t want to list the Chair in case some American Republicans found out they have one?

Of Mice and Men

Instead of statutory labor compliance and functioning governance, EFRAG’s report is packed with performative micro-metrics. It has several pages of meaningless environmental metrics. They don’t report on Scope 3 emissions. Or on the anticipated financial effects (AFE) of physical and transition risk from climate change in a 1.5°C scenario. The standards they have developed companies to require both. The AFE story is a rather shocking one and I’ll be writing more about that.

Management proudly discloses its efforts to reduce battery waste by favoring “wired devices (e.g. headphones, mice)”, encourages “black-and-white printing”, and precisely discloses that it generated exactly 1.25 kilograms (that’s 2.75 lbs. for my fellow ‘Mericans, the weight of a medium-sized Norway Rat) of hazardous waste printing toner. Well, here’s a Young Thug Whoopty Doo for the environment!

This clears neither the bar for financial materiality nor impact materiality. For a sustainability reporting standard-setter like EFRAG, their people, human capital, are what’s material in both a financial and impact sense. But, hey, remember that they’re paid above the minimum wage! And in a bold and defiant act of non-disclosure we have no idea what the total people costs are. We only know they have €16,715,700.80 in turnover. EFRAG’s impact on the world is entirely mediated through the staff drafting these sweeping regulations. How can an organization obsess over tracking a single kilogram of printer toner while ignoring a legal mandate to provide a safe grievance channel for its own workers?

While EFRAG operates as a Belgian non-profit association, they are essentially a subsidized arm of the European Commission. According to EFRAG’s own 2023 Annual Review, the European Union finances up to 90% of the eligible costs for their sustainability reporting work. One has to wonder what the Commission makes of this sustainability report, considering they have outsourced the drafting of their flagship standards and the credibility of the CSRD to an organization that tracks wired mice but cannot implement a basic statutory code of conduct.

Giving Some Tough Love

Tough love is some of the best love any of us can get. I realize I’ve given a rather heavy dose of that to EFRAG in this piece, but I am a long-term kind of guy. Standard-setters must have the institutional maturity to live under the principles they enforce, and I hope they can get there. Right now, EFRAG’s own reporting suggests they are struggling to manage their own house. I will be looking forward to seeing their 2026 sustainability report with great anticipation. Hopefully, next year, the dog food will taste a little better.

Oh, and before I go I’d like to give a little tough love to The Directorate-General for Financial Stability, Financial Services and Capital Markets Union (DG FISMA), the organ (weird term, I know, but again trying to be European 🐥) which oversees the work of EFRAG. Right now, EFRAG is doing a consultation for the Article 40a Standards which is how non-EU issuers will have to report on impact materiality. This will kick into effect for reporting for the year 2028. The consultation is open until the end of this month, and it is expected that the standards will be issued early next year.

Imagine what will happen if Republicans and the corporate community get a sniff about any irregularities in EFRAG’s standard setting process 😱. Saving on 1.25 kilos of hazardous wastepaper toner will look like turning off the water faucets before being hit by a Level 5 American hurricane. As that storm rolls in, DG FISMA shouldn't be surprised to find themselves echoing Ferdinand, the Prince of Naples, in The Tempest: "Hell is empty and all the devils are here."

Shiva Rajgopal and I recently published “Anticipating the Risk Factors in Anthropic’s IPO Prospectus: What they should be and what they will be.”

We are now running a follow-up survey, open through October 15, 2026. It asks you to rate 17 risks to Anthropic’s value, invites you to suggest others and add any comments, and asks how useful you find the risk factors in IPO prospectuses and 10-Ks. Responses are anonymous.

The timing matters. Anthropic’s prospectus has not yet been made public, so your ratings are a genuine forecast that we will be able to compare with what the company discloses. Shiva and I will publish the results in our respective Substack newsletters.

The survey takes only about five minutes, and we’d be very grateful if you would fill it out!

TAKE THE SURVEY

This piece is co-authored with Shivaram Rajgopal and also published on his Substack newsletter The Googly.

Anthropic has confidentially submitted a registration statement for what may be the largest initial public offering in history. Press reports put the valuation near $2 trillion and the proceeds as high as $100 billion, with trading as soon as November. When the prospectus is made public, readers will turn to the section called Risk Factors, and they will find dozens of pages of it. We want to put something on the record before they do: our own view of the risks that matter most to Anthropic’s value, and our prediction of how the prospectus will describe each of them.

We are doing this in advance, and dating it, for a simple reason. Once a prospectus is public, every analysis of it is shaped by it. The only way to learn whether careful work from public information can anticipate what a company will disclose, and what it will leave out, is to write the analysis down first and compare afterward. Everything here rests on information that was public on or before September 26, 2026. We completed the analysis before September 28, when Reuters and the Financial Times reported details from a version of the prospectus they had reviewed. Nothing from those reports has been incorporated, and none of our risks or predictions has been changed in light of them.

Our argument fits in one sentence. The risks that most threaten Anthropic’s value are not the AI-safety risks that dominate public debate about the company. They are a compute build-out of roughly half a trillion dollars in commitments set against revenue with only a few quarters of history; a supplier base that is also its shareholder base; and a federal government that twice in 2026 acted against the company by executive discretion. The safety and liability risks are real, but most of them are contingent and sit in the tail. Our prediction follows from how prospectuses are written. The document will be most specific about the risks the law requires it to describe, and least specific about the ones that drive value. One more thing belongs at the top. When our list was finished, one of us read it as an accountant and added two risks that rank with the most important here: the unit economics of selling AI, and the charge for stock-based compensation that will hit at the IPO. Our process missed both, and we explain below how.

How Risk Factors Actually Work

It is natural to assume that lawyers scrub a prospectus by leaving things out. For risk factors the pressure runs the other way. Under Section 11 of the Securities Act, an issuer is strictly liable for a material misstatement or omission in its registration statement. Its directors, the officers who sign it, and its underwriters can also be liable, although, unlike the issuer, they have a due-diligence defense. The safest course for counsel is to disclose everything that might conceivably matter, which is why technology prospectuses routinely carry 50 to 100 captioned risk factors. The SEC’s 2020 amendments to Item 105 of Regulation S-K responded to the resulting sprawl. They require a summary when the section runs past 15 pages, require headings, and discourage generic risks that could apply to any issuer.

The shaping happens in the manner of disclosure rather than the fact of it, and it takes four forms. The first is generality: a risk specific to the company is written in language that would fit any company in its industry. The second is hypothetical framing: an event that has already happened is described as something that could happen. The third is the absence of numbers: the risk is named and the magnitude is not. The fourth is burial: a serious risk sits among eighty others with nothing to mark it as different. Hypothetical framing is the one that has drawn enforcement. In 2019 Facebook paid the SEC $100 million for describing the misuse of user data as a hypothetical risk after it had learned that such misuse had occurred.

This tells us what is worth predicting. Almost every risk we identify will appear in some form. The informative question is how—named specifically or generically, as an event or as a possibility, with a number or without one. That is what our predictions address, and any reader can check them against the filing.

How We Arrived at Our List

We used nine industries whose histories of liability, reliance, and regulation offer a structural parallel to a frontier AI company: tobacco, asbestos, pharmaceuticals, social media, credit rating agencies, enterprise software, telecommunications, banking, and semiconductors. We did not use them as arguments that AI resembles any of them. Each industry was a device for generating questions. What did tobacco’s experience with internal knowledge and warnings suggest about Anthropic? What did the telecom capacity glut of 2000 to 2002 suggest about its compute commitments? What did the rating agencies’ conflict, in which the rated party pays, suggest about a company that grades the safety of its own models? We then tested the resulting risks against current public information, consolidated them, and only afterward compared them with the SASB standards. The Methodological Appendix describes the process. The full detail is in four working papers on Bob’s website: Working Paper 1, “Anthropic Before the Prospectus”; Working Paper 2, “Mapping the Inventory to the SASB Standards”; Working Paper 3, “Scoring Protocol”; and Working Paper 4, “Placebo and Null Arms.”

One disclosure belongs in the body of this piece rather than a footnote. The research and drafting were done with Claude, the AI model made by Anthropic. We chose to do it this way because whether an AI model can analyze its own developer is itself worth knowing. It also creates an obvious risk: output that looks like analysis may be recall of the company’s own self-description, which the model has seen in quantity. We guarded against this in three ways. Every factual claim had to trace to a dated public source. Anything the model could not trace was marked as unsourced in the working paper. And we ran the same method on a fictitious company, to see what it produces with no company-specific information at all. The appendix reports that test, including its limits, and its result is instructive.

The Risks We Think Matter

We identified 15 risks. Table 1 lists them with the fact that makes each material and our prediction of its treatment in the prospectus. We discuss them here in three groups.

Money and Infrastructure

The largest risk is compute. Anthropic has told investors it expects about five gigawatts of computing capacity by the end of 2026 and roughly double that by the end of 2027. According to reporting by The Information, it signed compute agreements totaling about $517 billion in the 11 months to September 2026. They include more than $100 billion to Amazon Web Services over 10 years; a $30 billion commitment to Microsoft Azure; up to a million of Google’s tensor processing units (TPUs), the custom chips Google designs for AI, plus further multi-gigawatt Google and Broadcom capacity from 2027; $50 billion of data centers with Fluidstack; $45 billion with Nscale; a twenty-year lease with TeraWulf; and $15 billion a year to SpaceX through May 2029. Much of the capacity is reportedly financed through special-purpose vehicles: more than $71 billion of chip-lease debt has been raised through them, and a $35 billion loan against data-center lease payments reportedly carries a Google backstop. Whether these obligations appear on Anthropic’s balance sheet is one of the things the prospectus will settle.

The revenue these commitments serve has grown at extraordinary speed, from about $9 billion annualized at the end of 2025 to more than $65 billion in July 2026. But first-quarter 2026 revenue was about $4.8 billion, and the company told investors its first operating profit, projected for the second quarter, might not last the year. Twenty-year obligations are being set against a revenue history measured in quarters. The mismatch is one of duration as well as scale. The obligations are long-dated and largely fixed, while much of the revenue is usage-based and short-term. A fall in token prices, the loss of a large customer or tighter capital markets could leave Anthropic paying for capacity its operations cannot finance, and dependent on raising new capital precisely when it is hardest to raise. Having enough capacity and being able to keep paying for it are different problems. The telecom build-out of 1996 to 2000 is the cautionary parallel, when capital spending doubled ahead of demand and the sector then lost about $700 billion of market value. The parallel cuts both ways, however, because Anthropic’s visible problem today is shortage: its outage record and two consumer class actions over usage limits are consistent with demand outrunning capacity, though they do not prove it. Energy is the physical constraint underneath all of this. Record capacity prices at PJM Interconnection, the operator of the power grid serving 13 mid-Atlantic and Midwestern states and Washington, D.C., have made data-center demand a political issue, and Anthropic has pledged to pay for the grid upgrades its data centers require and to cover consumer price effects.

The second risk is who supplies that compute. Anthropic’s principal suppliers are also its principal investors. Amazon’s investment facility is tied to compute-delivery milestones, and its Anthropic interests are carried at about $190 billion. Google has committed up to $40 billion of investment while supplying TPUs and guaranteeing lease financing. Microsoft and Nvidia invested alongside the Azure commitment. AMD’s equity investment accompanies its chip supply, and Broadcom backs the lease vehicles. Amazon, for one, books large gains as Anthropic’s valuation rises, while Anthropic’s spending becomes its suppliers’ cloud revenue as the services are delivered. Whatever the commercial logic, this is a concentration of counterparty, pricing, and governance exposure that equity analysts will want to see set out plainly. The third risk is the other side of the ledger: revenue concentration. In 2025, two coding products, Cursor and GitHub Copilot, were reported to account for nearly a quarter of Anthropic’s revenue, and Microsoft is at once a channel, a supplier, an investor, and a competitor.

The last item in this group is not a risk in the usual sense but a disclosure question: how revenue is measured. Anthropic’s growth has been communicated almost entirely through annualized run-rates. A prospectus will present revenue under GAAP, and the reconciliation between the two will matter. So will a less discussed accounting question. When Claude is sold through Amazon’s and Google’s cloud marketplaces, whether Anthropic records the gross sale or only its net share. With investors who are also resellers, the answer affects both reported growth and the reading of the related-party relationships.

Government

Two events in 2026 showed that Anthropic’s access to its own markets depends on executive discretion. On June 12 the U.S. government issued an export-control directive suspending access to Anthropic’s two most capable models, Fable 5 and Mythos 5, by any foreign national. The concern cited was a technique for getting the model to read code and fix software flaws, which is the capability Anthropic sells to cyber defenders. Anthropic disabled both models for all customers to comply. Access was restored on July 1, on conditions that included reporting malicious activity to the government. Nothing was priced and nothing was litigated. Permission was simply withdrawn and then returned. Anthropic advocates strict export controls on AI chips while having contested the control applied to its own models. The two positions are distinguishable in principle, but both coincide with its competitive interest.

The second event is still live. In February Anthropic refused to allow its models to be used for mass domestic surveillance or fully autonomous weapons, and the Department of War acted against the company under two separate statutes. The Secretary stated publicly that no contractor doing business with the military may conduct commercial activity with Anthropic. In August a federal district court invalidated the supply-chain-risk designation made under the first, 10 U.S.C. § 3252. On September 25, the day before our cutoff, the D.C. Circuit upheld the exclusion made under the second, the Federal Acquisition Supply Chain Security Act, by two votes to one. Anthropic reads the restrictions narrowly, as covering only use of Claude as a direct part of defense contracts. The financial stakes depend almost entirely on how far the restrictions ultimately reach, because Anthropic’s largest suppliers are themselves defense contractors. Few disclosures in the prospectus will be read more closely.

Liability and Governance

The liability that has already crystallized concerns training data. Anthropic settled Bartz v. Anthropic for $1.5 billion after the court held that training on lawfully acquired books was fair use but that downloading and keeping pirated copies was not. By August $1.05 billion had been paid into the settlement fund. The final $450 million is reported to fall due in September 2027 or 30 days after a “qualifying funding event,” whichever is earlier. An IPO may be such an event, and the filings we reviewed do not define it. The settlement covered past acquisition only. Three music-publisher suits followed. The second relies expressly on evidence surfaced in Bartz and names two of Anthropic’s co-founders personally: Dario Amodei, the company’s chief executive, and Benjamin Mann. That suit lists about 21,000 works. At the statutory maximum it alone could reach about $3.2 billion; at the Bartz rate of about $3,000 per work it would be about $64 million. Under US accounting, a loss that is probable and can be reasonably estimated must be accrued, at the low end of the range if no amount within it is a better estimate; otherwise it is disclosed. The width of this range shows how uncertain the exposure is, and the prospectus will show which course the company has taken. Anthropic also indemnifies its business customers against intellectual-property claims arising from their use of Claude, outside its ordinary contractual limit of liability.

The liability that has not crystallized concerns harm from use. We found no publicly reported wrongful-death or product-liability suit against Anthropic as of our cutoff. The comparators are nonetheless close. Courts in cases against Character.AI and OpenAI have declined at the pleading stage to treat chatbot output as protected speech. Whether Section 230 of the Communications Decency Act of 1996, which shields online platforms from liability for content posted by their users, protects a developer from claims over content its own model generates remains unsettled, and the protection is weakest where the developer is alleged to have helped create the content rather than merely transmit it. Meta’s settlement with state attorneys general over harm to minors, reached with a bipartisan coalition of forty-seven states and the District of Columbia, is worth up to $17.1 billion and mandates age verification. Anthropic received both of the attorneys-general letters sent to AI companies in 2025, which demanded safety testing, recall procedures and warnings. Two further risks sit alongside harm from use. One is misuse: Anthropic has itself disclosed a state-sponsored espionage campaign in which its model did most of the work, and insurers have reportedly balked at covering the laboratories’ exposures. The other is agentic products, which act in customers’ systems, so that a failure takes the form of an action rather than an answer. Anthropic’s standard contracts generally cap its liability at a year’s fees, and we found no reported decision testing that cap against an agent-caused loss.

Regulation is multiplying quickly. California’s SB 53, which Anthropic endorsed, requires frontier developers to publish safety frameworks and report critical incidents. New York’s RAISE Act, which takes effect on January 1, 2027, will impose a seventy-two-hour reporting clock. The EU AI Act‘s obligations for general-purpose models have applied since August 2025; the European Commission’s power to enforce them began in August 2026, and models already on the market have until August 2027 to comply. Anthropic has given $40 million to Public First Action, a group that backs candidates and a platform favoring transparency legislation. We found no public position from Anthropic on the question that matters most financially: who should bear liability when a model causes harm.

Anthropic’s own safety commitments are a risk in a sense that is unusual for a prospectus. They can require the company to forgo revenue. Its Responsible Scaling Policy can require it to delay or restrict a model. It has withheld its most capable cyber model from general release. Its August risk report disclosed an internal model that it does not plan to release. The policy was also revised in February 2026 in a way that removed an earlier commitment to pause development. Jared Kaplan, a co-founder and Anthropic’s chief science officer, explained that unilateral commitments did not make sense “if competitors are blazing ahead.” Investors will want to know which commitments bind and who can change them.

Finally, governance. The founders are reported to be seeking shares carrying 50.1 percent of the vote on most matters. The seven co-founders own about 2 percent of the company apiece, and the extra votes carry no economic value. Investors have accepted the separation of control from ownership before. Mark Zuckerberg holds 61 percent of Meta’s vote on 13 percent of its economic ownership, and under the registration statement SpaceX filed in May 2026, Elon Musk holds 85.1 percent of its vote. Anthropic goes further in one respect: control is split between founders with small stakes and a Long-Term Benefit Trust with none, which selects a majority of the board. The company is also a Delaware public benefit corporation, obliged to balance shareholders’ financial interests, the interests of those materially affected by its conduct, and the public benefit named in its charter. SaferAI, a nonprofit that rates the risk management of AI developers, found in July 2026 that Anthropic has a designated Responsible Scaling Officer but no central risk function and no management-level risk committee. The banking parallel is instructive. Banks also depend on models they build themselves, and large U.S. banking organizations are required to have an independent chief risk officer and a board-level risk committee. Anthropic faces no such requirement and has no such structure.

What We Expect the Prospectus to Say

Our predictions for all 15 risks are in Table 1. The pattern is what matters. We expect the prospectus to be specific wherever the law or the facts leave no choice. The copyright litigation will be described in detail, because pending legal proceedings must be. The June export suspension will be described as an event, because describing it as a possibility would invite the Facebook problem. The founder share class, the Trust and the public-benefit structure will be set out at length, because governance of that kind must be. The Pentagon designation will be disclosed, but we expect it to be read narrowly. Regulation will get extensive and accurate treatment.

We expect the prospectus to be least specific where the company has discretion and the stakes are highest. The compute commitments will appear as numbers in the notes to the financial statements and the discussion of liquidity; SEC rules no longer require a separate table of contractual obligations. The risk around them will be framed as needing more capacity, not as having committed to too much. The need for further capital will be presented as a conventional growth-company risk, not as a question of whether financing will be available when weaker demand, falling margins or a lower valuation make it most necessary. The supplier-investor relationships will appear as related-party transactions, as they must, without being characterized as a concentration of risk. Customer concentration will be anonymized. Harm from use and misuse will be described in general terms, with insurance limits unquantified. We do not expect the company to state a position on liability allocation. Nor do we expect it to quantify the revenue its safety commitments have cost it, or to say whether the offering accelerates the final Bartz payment.

We also expect the prospectus to give heavy weight to risks our analysis did not rank highly. These are competition among frontier laboratories, commoditization by open-weight models, the rapid obsolescence of each model generation, dependence on key personnel, and the security of model weights and source code (Table 2). We say so now so that their prominence is not mistaken for a miss on our part. Our industry analogies were chosen to surface liability, reliance and control risks, and we knew from the start that they would not generate commercial ones. As the next section shows, that boundary also cost us two risks that belong among the most important.

When the prospectus is public we will compare it against this piece and report what we got right, what we got wrong, and what the filing told us that public information could not. For anyone who wants to make that comparison systematically, Working Paper 3 provides a scoring protocol. We will also return to the question we have deliberately set aside here: what these risks are worth in dollars. That analysis is better done with the audited numbers in hand.

What a Domain Expert Added

The industry analogies were Shiva’s idea. Bob then built the list out with Claude under the process described above. When the 15 risks, the predictions and the working papers were finished, Shiva read the result as what he is, an accounting professor who reads financial statements for a living, and wrote down 10 risks of his own. Table 3 sets them against ours. We present them separately, and leave Table 1 as it was, because the point of this section is to show what a domain expert added after a machine-assisted process had run its course. The working papers predate his review and are unchanged.

Two of his items are the most important thing our process missed, and they are not details. The first is unit economics: whether the price of a token falls faster than the cost of serving it. It combines his first item, cheap Chinese models, and his fourth, token costs that are not falling fast enough for mass adoption. In late September 2026, published price lists put DeepSeek’s V4 Pro at roughly a tenth of the input price and a fifth of the output price of Claude Sonnet 5, Anthropic’s mid-tier model, which charges $2 per million input tokens and $10 per million output tokens. These prices change often, so the comparison is a dated snapshot rather than a fixed ratio. On the cost side, Anthropic was reported early this year to have cut its 2025 gross-margin projection to 40 percent, with inference costs running above plan. Those facts were in our own source register, and we did not turn them into a risk. Unit economics is arguably the most important driver of Anthropic’s value after the compute commitments, and it is the mechanism that would turn those commitments into losses. Table 2 declared price competition out of scope because our analogies were not built to find commercial risks. We now think that was a convenient way of excusing a gap rather than a principled boundary.

The second is stock-based compensation. The large charge at an IPO does not come from revaluing awards at the offering price; under US accounting they are valued when granted. It comes from restricted stock units that vest only on a liquidity event. Until the event occurs no expense is recognized, so years of accumulated compensation cost land in the first public quarter. Facebook is the precedent. In the quarter of its 2012 offering it recognized $1.3 billion of share-based compensation and related payroll taxes and reported a GAAP net loss of $157 million. Whether Anthropic’s awards are structured this way is not public, and the prospectus will say. We predict a specific, quantified disclosure, a large GAAP loss in the first reported quarter, and dilution that analysts will need to model.

Three of Shiva’s items, years without profit, perpetual cash burn, and the compensation charge, are one risk seen from different sides—the price of patience. Amazon is the obvious precedent, and the instructive thing about it is that it is the same issue in a very different economic environment. Jeff Bezos told shareholders in his 1997 letter that when forced to choose between “optimizing the appearance of our GAAP accounting and maximizing the present value of future cash flows, we’ll take the cash flows.” Investors accepted that, and the capital to act on it was cheap. In January 1999 Amazon raised $1.25 billion in convertible notes paying 4.75 percent. Much of its pay took the form of options, whose cost was reported only in the footnotes until accounting rules required it to be expensed from 2006. Interest rates stayed low for most of the two decades that followed.

Anthropic is asking for the same patience on a far larger scale and at a far higher price. The junior tranche of its chip-lease debt, which lacks a Broadcom guarantee, was reported in August to trade at about 8.5 percent. Its commitments run to hundreds of billions of dollars, not $1.25 billion, and its compensation cost will be visible from the first quarter. We add a prediction: the prospectus will include a letter from the founders, in the Amazon tradition, asking investors to judge the company over the long term. What the letter will not settle is whether investors will extend Amazon’s patience at today’s cost of capital. We read items (vi) and (vii) as part of risk 1, which should be understood as the compute commitments and their financing.

Shiva’s point about distribution sharpens one of ours. We said that Microsoft is at once channel, supplier, investor, and competitor. He points out that Microsoft and Google are powerful gatekeepers to corporate and individual customers, which leaves Anthropic dependent on intermediaries that also compete with it. That is the better sentence.

He also raised a question we had not asked. The leaders of the frontier laboratories, Amodei among them, describe the risks of AI in near-apocalyptic terms. Is that an attempt to become too big to fail? We cannot know the motive and do not speculate about it. The effect is easier to see, and it bears on misuse (risk 11). When a company says publicly and repeatedly that its products could contribute to catastrophic harm, those statements can help a plaintiff establish that the harm was foreseeable, and they will matter to any insurer pricing the risk. Tobacco’s liability rested on internal documents showing that the companies knew what they denied in public. Here the knowledge is published. Insurers have already reportedly balked at covering the laboratories’ largest exposures. The likelier outcome is not too big to fail but too big to insure, in which the company retains the tail risk itself or, in the extreme, the government does.

The rest of his list adds less. Data-center delays, years of losses and the public benefit corporation are in our fifteen. Key-person risk is in Table 2; Amodei has himself written about surviving early-stage cancer, and we expect the prospectus to name the founders as key persons without reference to health. Nationalization abroad does not survive contact with the facts, since Anthropic has little outside the United States to take. The realistic versions are U.S. compulsion and foreign market-access rules, which are tails of risks 6 and 7.

The pattern is the finding. Shiva’s list is an investor’s and an accountant’s list: the income statement, the cash-flow statement, dilution, and competition. Ours is a liability, regulation, and governance list. His items came quickly from expertise; ours came from a heavy, sourced process. His list will probably predict the prospectus’s own risk factors better than ours, because counsel and bankers think in his categories. Ours is better at identifying what the prospectus will underplay. Nine industry analogies, a disciplined sourcing process and an AI model with the relevant facts in its own register did not produce the two risks an accountant produced by reading the result. The honest description is not that the human refined the machine’s work at the margin. The machine-assisted process was thorough within the frame it was given, and the human saw outside the frame, in the direction that matters most to an equity investor.

Methodological Appendix

Sources and Cutoff

Every factual claim in this piece and in the working papers rests on a source that was public and dated on or before September 26, 2026. The sources are listed in a register of 151 entries in Working Paper 1 each giving the claim, the source, the URL, the publication date and the type of source: company statement, regulatory filing, court document, government source, press report, or analyst estimate. As of the cutoff, Anthropic’s public registration statement had not appeared on EDGAR; the company had announced a confidential submission on June 1. Press-reported figures, especially revenue run-rates and the aggregate of compute commitments, are not audited, and we treat them as such.

The Industry Analogies

We chose nine industries which we then organized into three groups. The first group concerns harm and liability arising from use: tobacco, asbestos, pharmaceuticals, and social media. The second concerns reliance, intermediation, and the contractual allocation of liability: credit rating agencies, enterprise software and telecommunications. The third concerns systemic exposure and control regimes: banking and semiconductors. For each industry we identified the structural feature doing the work: tobacco’s internal knowledge and warnings, asbestos’s indeterminate and ultimately uninsurable tail, pharmaceuticals’ regime of pre-approval testing and post-market surveillance, social media’s erosion of intermediary immunity, the rating agencies’ conflicted payment model, enterprise software’s contractual allocation of implementation failure, telecom’s capacity glut and common-carrier obligations, banking’s independent risk function, and semiconductors’ concentration and export controls. We then asked what each feature implied for Anthropic, and we kept only risks we could anchor in current public facts. Table A1 lists the industries and what each contributed. The analogies are generative, not probative. None of them is an argument that AI is like tobacco or asbestos. Their value is that they surface questions a reader of the company’s own materials would not think to ask.

The analogies produced overlapping risks, which we consolidated into the fifteen in Table 1. Where several industries arrived at the same risk by different routes, we treated the convergence as evidence that the risk was real rather than an artifact of any one analogy. The copyright exposure is the clearest case. It was reached separately through tobacco (what executives knew), asbestos (the reuse of evidence by successive plaintiffs) and social media (discovery), and all three routes are visible in the second music-publisher complaint.

The SASB Comparison

We deliberately did not consult the SASB standards until the list was complete, so that the existing catalog would not shape it. SASB identifies the sustainability issues likely to be financially material in each industry. It was never meant to identify every material issue, and competition, financing, customer concentration and accounting are deliberately outside it. So we asked two questions of each risk, including Shiva’s two additions. Is it a sustainability issue in SASB’s own terms, falling within one of its twenty-six general issue categories? And if so, is it a disclosure topic in the standard for Software & IT Services, the industry to which Anthropic would be assigned? Of the 17 risks, seven are clearly sustainability issues, three are partly so, and seven are not. The seven that are not (the compute commitments, the supplier-investors, customer concentration, revenue measurement, the IP indemnity, unit economics and stock compensation) are outside SASB’s scope by design, and they include the largest risks in dollar terms. Of the 10 that are sustainability issues, only one, energy and data-center infrastructure, is a topic in the Software & IT Services standard. Four more fit only with strain. The other five, among them harm from use, misuse, frontier-AI regulation, and the cost of safety-driven restraint, sit squarely within SASB’s remit and are absent from the standard that would apply. They are product-safety risks of a kind the standards handle in pharmaceuticals, tobacco, and consumer goods but not in technology. The finding is timely. In July 2026 the International Sustainability Standards Board decided to enhance all six SASB standards in the Technology & Communications sector. Working Paper 2 maps the original 15 risks topic by topic; the two-question frame used here is a refinement of it.

Working with Anthropic’s Own Model

The research, drafting, and source-gathering were done with Claude, working under our direction. The model had no non-public information about Anthropic. Because it was trained on a large volume of Anthropic’s published self-description, we required every factual claim to trace to a retrieved and dated source. Any claim it could not trace was marked in line as unsourced, and Working Paper 1 shows those markings. The risks most exposed to this problem are the ones that depend on Anthropic’s descriptions of its own safety governance, and we have flagged them there. A separate instance of the model, which had not taken part in producing the inventory, checked every cited claim against the source register and re-fetched the highest-stakes sources; its corrections were made before publication.

We also tested what the method produces when the company-specific information is thin or absent. These tests were run inside our working session with Claude, not in separate conversations. Claude launched each as a sub-agent that began with an empty context: the sub-agent could not see our conversation, the Anthropic inventory or the purpose of the test, and it received only the brief used for Anthropic with the company name changed. But the briefs were written, and the results interpreted, by the same session that produced the Anthropic analysis. The tests are therefore not independent of us, and they should be read as a check on the method rather than a controlled experiment.

The first test used Safe Superintelligence, a thinly documented laboratory. By September 2026 it had enough public substance, including a large Nvidia investment in July, to yield a real analysis, so it could not serve as a placebo. The second used a fictitious laboratory whose name returned no match in any search. The model invented no facts about the fictitious company; it stated at the outset that no public record existed. But the method still produced a fourteen-item inventory of risks for frontier laboratories as a class, and that inventory reproduced eight of our 15 Anthropic risks in full and three in part. We draw a clear lesson. Much of what an analysis like ours says about Anthropic would be said about any frontier laboratory.

The company-specific value lies in four risks the generic inventory did not produce (customer concentration, the Pentagon designation, energy, and revenue measurement) and in the evidence attached to every risk. That evidence includes the Bartz payment terms, the June suspension, the supplier-investor web and the founders’ proposed share class. It is also where our predictions should be judged. A further caution; the model’s picture of frontier-laboratory risk in general was learned substantially from Anthropic’s own public record, so agreement at the generic level proves little. The generic run retrieved the April 2026 revision of bank supervisors’ model-risk guidance, which excludes generative and agentic AI and which our own inventory had initially missed and has since corrected. Working Paper 4 reports both tests.

Limitations

The main limitations are four. First, the figures are press-reported rather than audited, and some sources conflict; the register records the conflicts. Second, the analogy set was chosen for liability and control risks, and it under-represents commercial and accounting risk by design; the section on what a domain expert added shows what that cost. Third, the placebo and null tests were run once each, from within our own working session, by the same model; they are suggestive rather than conclusive. Fourth, a stronger version of those tests would be run in a separate conversation by someone other than us, and with web search disabled.

Dan Crowley, a lawyer at K&L Gates who focuses on financial services, has been a good friend of mine for over 20 years. And he’s a Republican. All of you know I’m a Democrat. Not surprisingly, there are issues on which Dan and I disagree. But there are far more issues we agree on than you might think.

One thing we agree on is that the ESG Culture Wars have been silly and damaging to America’s capital markets. Our first piece was published in the Harvard Law School Forum on Corporate Governance (on September 1, 2022. In “Turning Down the Heat in the ESG Debate: Separating Material Risk Disclosures from Salient Political Issues” we focused on climate and argued that it was up to the SEC to determine what material issues should be disclosed. I am disappointed with the Commission’s decision to rescind the climate-related disclosure rules although Dan has told there are questions about its authority to do so. That is a story for another day.

Our next piece on this topic, “Rescuing ESG from the Culture Wars” was published in the Harvard Business Review on February 9, 2023. Our final sentence: “To us, ESG is simply about identifying material risk factors that matter to company profitability and shareholder value over time.”

Our third piece, “Here We Go Again: Red States Continue to Focus on ESG” was published in the Harvard Law School Blog on September 4, 2025. This piece was about the July 29, 2025 letters from 26 members in 20 red states and the purple state of Pennsylvania of the State Financial Officers Foundation (SFOF) to the leaders of 25 asset managers. Somewhat ironically the Mission of the SFOF is “to drive fiscally sound public policy by partnering with key stakeholders and educating Americans on the role of responsible financial management in a free market economy.” Dan and I argued that investors who ignore ESG risks are more likely to be in violation of their fiduciary duty than those who do not.

Making it a quartet (albeit a shorter one than Lawrence Durrell’s famous “The Alexandria Quartet”) on September 17 the Harvard Law School Blog published our analysis , “The Red State Attack on ESG Continues to Misfire,” of the letter from 16 red state Attorneys General to the Big Four accounting firms. We conclude that “In short, the assertion in the AGs’ letter upon which all alleged legal violations rely, do not withstand scrutiny. It will certainly generate coverage, though we urge caution against anyone who may wish to build on this alarming precedent.”

Four years on and speaking for myself, I’d be happy to call it a day with a quartet. But my instinct is that there will be future opportunities for Dan and me to write about ESG together.

16 September 2026

I have been a big supporter of reporting standards for sustainability performance for nearly 40 years. I was the Founding Chairman of the Sustainability Accounting Standards Board (SASB). With Richard Barker, then a colleague of mine at the Saïd Business School at the University of Oxford and now on the board of the International Sustainability Standards Board (ISSB), in 2018 I wrote a Green Paper titled “Should FASB and the IASB Be Responsible for Set Standards for Nonfinancial Information?” Three years later at COP26 the IFRS Foundation announced the launch of the ISSB. Richard and I were so excited! We were also excited when Sue Lloyd, who had a distinguished career at the International Accounting Standards Board (IASB), was named the Vice-Chair. This made perfect sense. Sue has deep expertise in financial accounting and understands the importance of having quality standards for sustainability reporting to supplement financial statements and help explain financial performance. She and I have been in touch over the years and more so recently as the debate about ESG and sustainability rages on, particularly in the U.S. Some of this is due to people simply not understanding the purpose of the ISSB. I figured now was a good time to help make clear to everyone the purpose of the ISSB and how it does its work. I asked Sue if she would be willing to do an interview with me and she kindly agreed.

Eccles: Sue, thanks for taking the time for this conversation. For more than 20 years you have been setting standards to help companies report information about their financials, but you never worked as an accountant. Is this not surprising?

Lloyd: It is certainly surprising for professors of accounting! When I was a teenager in New Zealand, I wanted to be an accountant, but after my degree in accounting and finance I started my career in investment banking rather than accounting.

Eccles: It sounds like you were taking Monopoly way too seriously when you were a child! Did you have any childhood habits that hinted you might end up in international accounting?

Lloyd: I am not sure that as a child anyone would have thought I would end up working in international standard setting and being based outside New Zealand for my entire career. I grew up in a small country town in New Zealand. My upbringing was very much a rural one. My holiday jobs were pruning and picking kiwi fruit. I didn’t even have a passport until I needed one for the job interview in Australia that led to me moving to Sydney in my early 20s. People probably wouldn’t be surprised to hear that I have always been pretty academic and studious. You would have been more likely to find me playing the piano or studying than on a sports field. Perhaps the only sign of the international life to come was that my thesis for my Masters in Accounting and Finance at Auckland University was about harmonisation of accounting internationally.

Eccles: How did that job interview in Australia go?

Lloyd: It seems well. It led me to pack up and leave New Zealand – at that time with the clear intention that it was for two years. That plan didn’t quite pan out – I won’t say how many years later having worked overseas since, but that ended up being the direction of my career. In investment banking, I worked in structured debt teams initially and later in structured equity derivatives and dealt extensively with accounting standards. I often found myself frustrated by the difficulty of using them to reflect the economic reality of transactions. My interest in standard-setting grew out of a desire to improve the way transactions were represented in financial reporting so that investors and other stakeholders could better understand what was really going on.

Eccles: Which accounting standards were you dealing with at that time?

Lloyd: I worked in Hong Kong, Sydney and London, but started out first working in Australia. My first role in standard setting was as a member of the Australian Accounting Standards Board (AASB), so mainly Australian accounting standards, whose conceptual foundations are very aligned with what are now IFRS Accounting Standards. I was actually on the AASB when Australia made the decision to move to IFRS Accounting Standards. And showing my age here, I remember getting faxed copies of international accounting standards from a central library source in Australia as they started to become increasingly relevant in the late 1990s.

Eccles: From Auckland to London, via Hong Kong and Sydney. What attracted you to the IASB and to leave what must have been a profitable job in investment banking?

Lloyd: The possibility to make a difference. I hoped that I could use my market experience to help ensure IFRS Accounting Standards were high quality. I’m also genuinely interested in standard setting. It is something I am passionate about. I strongly believe that high-quality standards matter. They are fundamental to the effective operation of capital markets.

Eccles: When did your journey with the IASB begin and what was it like working there at that time?

Lloyd: I have been fortunate to be at the IASB at some important moments in its history. I was there as a staff member in the early 2000s in the early days of the IASB when David Tweedie was Chair, when the IASB was getting its standards ready for adoption in the European Union. I went back to banking for a few years and then came back again to the staff in 2009 after the global financial crisis, which was an exciting time in standard setting.

Eccles: Why was it so exciting?

Lloyd: I was fortunate to work on some very significant projects that transformed accounting internationally. I worked closely with IASB Chair Hans Hoogervorst from 2011 to 2021, as a senior technical director and then ultimately as Vice-Chair. We completed IFRS 15 on revenue recognition, IFRS 9 on financial instruments, IFRS 16 on lease accounting and IFRS 17 on insurance contracts. Each addressed areas where confidence in reported information had been weakened and where investors needed better information.

Eccles: The so-called Big4 IFRS Accounting Standards. Which one do you consider particularly transformative?

Lloyd: I am sometimes called the Mother of IFRS 9 so it is appropriate I choose that! It brought accounting closer to the way credit risk is actually managed. It was fundamentally about ensuring that financial statements provided a more relevant picture of risk and reflected economic reality more effectively and on a timely basis. The IASB had already started a complete overhaul of financial instruments reporting at the time of the financial crisis of 2008, but the crisis reiterated the need for transparency. The expected credit loss model in IFRS 9 was a critical response to this and close to my banking heart of course.

Eccles: While juggling your IASB Vice-Chair responsibilities, you were instrumental in bringing the ISSB to life, including driving the development of the ISSB’s two inaugural Standards. Given your unique perspective, how do you view the relationship between financial statements and sustainability-related disclosures?

Lloyd: The IASB has always been focused on providing information to enable investors to make investment decisions. Over time, the range of information investors need to understand the prospects of companies they are investing in has broadened. Sustainability-related disclosures fill that gap.

Eccles: How?

Lloyd: Sustainability-related disclosures provide additional information that investors need to understand a company’s current and potential financial performance. So, sustainability-related financial disclosures complement the information in the financial statements, providing more context about how business performance has been affected by things like climate-related risks and also the anticipated effects of those risks and opportunities. When applying ISSB Standards, disclosures are presented in a way that enables investors to understand how the information provided about climate and other risks and opportunities relates to information in the financial statements. So, looking at the information together, investors should see a cohesive story that enhances their understanding of a company’s performance and prospects.

Eccles: Can you give me an example?

Lloyd: It may be the case that a company has plans to make changes to its business, for example to improve energy efficiency, that will involve future investment. That may not be recognised in the financial statements yet given accounting requirements. The sustainability-related disclosures should enable those complementary perspectives to be understood.

Eccles: There’s been a growing narrative that ESG reporting is facing a backlash—with some even declaring it ‘dead.’ From where you sit, what is the reality on the ground and what would you say to the sceptics?

Lloyd: What we are seeing is the evolution of ESG reporting in sustainability reporting. The market is moving away from vague, disconnected, fragmented ‘ESG’ commentary towards decision-useful financial disclosures. First reports prepared in accordance with ISSB Standards are not too long and focus on reporting information that really matters for investment decision-making.

Eccles: Where is this reporting coming from?

Lloyd: More than 45 jurisdictions—representing over half of global GDP—are already taking steps to adopt or use the ISSB Standards in their regulatory frameworks. Reporting is starting in accordance with regulation in some jurisdictions, but even ahead of regulations kicking in, companies around the world are choosing to adopt these Standards on a purely voluntary basis.

Eccles: Why?

Lloyd: Because investors are demanding clear, comparable data on sustainability-related risks and opportunities, and business leaders recognise the value of high-quality disclosure to communicate efficiently and effectively with providers of capital. Applied well, ISSB Standards allow companies to provide information about matters that truly drive value creation and to explain to investors their resilience to the risks that they face. So, to the sceptics, I would say to look at the market fundamentals. Capital markets continue to need information about what matters. Disclosures help capital to move where it can be most productive.

Eccles: Critics often question whether standard setters are driving a specific sustainability agenda or exist to promote specific environmental or climate transition policies.

Lloyd: Let me be very, very clear that the ISSB is policy neutral. We do not require companies to set targets or to develop a plan to reduce GHG emissions. We do not tell companies how to run their businesses. We do not advocate for specific climate goals, such as achieving net zero by 2050. We do not evaluate whether a company’s strategy is good or bad.

Eccles: So, what are ISSB Standards about?

Lloyd: ISSB Standards are all about risks and opportunities. They require companies to disclose information about the climate and other sustainability-related risks and opportunities they have identified that could truly affect their financial prospects. And to provide information about what, if anything, they are doing about them and what that means for the company’s strategy and business model.

Eccles: Do you have an example in mind?

Lloyd: Does your company need to access a significant volume of water to run its agricultural product business? Are there issues with access to water due to this being in water-stressed areas? How are you dealing with this risk of water scarcity? That’s what the ISSB Standards are about. It’s business information, not about old style ESG. People also think climate reporting is just about carbon emissions and it’s not. It’s about how the business may need to change over the short, medium and long term in response to climate risks and opportunities, including physical climate risks. The aim is to enable investors to have reliable information about companies’ exposures, plans and prospects to decide how to allocate capital to companies.

Eccles: The ISSB is not affiliated with any jurisdiction, but you did not start from scratch, right?

Lloyd: We have not re-invented the wheel. We intentionally built on existing, capital markets-oriented frameworks and standards to ensure that we were using concepts and practices that were well known to market participants. The creation of the ISSB involved the consolidation of the Climate Disclosure Standards Board (CDSB) and the Value Reporting Foundation, which governed materials from the International Integrated Reporting Council (IIRC) and from the Sustainability Accounting Standards Board (SASB). In developing the ISSB Standards, we also built on the work of the IASB. For example, the definition of material information in ISSB Standards is the same as in IFRS Accounting Standards.

Eccles: The roots of the SASB lie in the U.S., which I know well having been its Founding Chairman. So, the SASB material was born in the U.S. and it is now being internationalised?

Lloyd: Our standards are really about what is material information for the particular entity. We think the industry-based materials developed by the SASB support this by enabling companies to provide industry-specific disclosures across the range of sustainability matters that are relevant to their business activities. And industry-specific disclosures enable better comparisons between companies, which is aligned with how investors conduct research and make investment decisions. And when you look at the statistics of who use the SASB Standards you see that 70% of companies using the SASB Standards are outside the U.S.

Eccles: So, you have incorporated these materials in the ISSB Standards as they are already used internationally?

Lloyd: We reference the SASB Standards as guidance to assist companies to provide information that is relevant for investment decision-making. Reflecting the international focus of the ISSB we had a dedicated project to ‘internationalise the SASB Standards’, for example by ensuring some specific U.S. references are more relevant to an international setting so that the SASB materials could be used internationally.

Eccles: You naturally share a close link with the IASB as you operate under the same IFRS Foundation umbrella, but when talking about financial reporting do you see anything in common with the U.S. Financial Accounting Standards Board (FASB)?

Lloyd: If you compare the definition of material information under IFRS Accounting Standards or ISSB Standards with the definition of material information under US GAAP you will see it is conceptually identical, focusing on whether an item’s omission or misstatement could influence economic decisions made by investors.

ISSB/IASB: “Information is material if omitting, misstating or obscuring it could reasonably be expected to influence investor decisions.”

FASB: “Information is material if there is a substantial likelihood that a reasonable investor would view the omission or misstatement as having significantly altered the total mix of information made available”

Eccles: We all take it for granted that we had those financial accounting standards, but we did not have anything in the U.S. until the SEC was formed. We had country gap for a long time.

Lloyd: In consolidating the global sustainability disclosure landscape, we are following the harmonisation path that the IASB took two decades ago. In the early 2000s, companies faced a patchwork of national accounting standards (such as UK GAAP, French GAAP, US GAAP), making international corporate performance incredibly difficult for global investors to compare. With IFRS Accounting Standards now used in more than 140 jurisdictions, the world’s capital markets are fundamentally dependent upon a single set of standards so investors can compare information developed using the same set of standards – in other words, people can do ‘apples to apples’ comparisons.

Eccles: So, are you leveraging the IASB’s experience?

Lloyd: What we are doing for the ISSB is the same thing but for a different set of investor-focused information that is complementary to the financial statements, using the same definition of material information. Companies want to just have to report according to one set of standards. Investors want to get information according to one set of standards so they don’t have to make the adjustments themselves to get to comparable information. That’s why we’re working for a global baseline of sustainability disclosures.

Eccles: How did you decide to start this journey to develop a global baseline?

Lloyd: We asked a question. Well, the trustees of the IFRS Foundation asked a question, via a consultation they launched in 2020: “Is there a need for a global set of internationally recognised sustainability reporting standards? If yes, should the IFRS Foundation play a role in setting these standards and expand its standard-setting activities into this area? If not, what approach should be adopted?”

Eccles: And I guess the response was that the IFRS Foundation had to play a role?

Lloyd: The feedback from the consultation made it clear that there was an urgent, global demand for comparable sustainability-related information for investment decision-making. Stakeholders, including the G20, G7, IOSCO and Financial Stability Board, called on the IFRS Foundation to take a leading role in setting international sustainability disclosure standards. The summary of feedback was published in April 2021, when I was still the IASB Vice-Chair. One year later we had the Exposure Drafts of IFRS S1 and IFRS S2 out for consultation and I was the ISSB Vice-Chair. The call to consolidate and reduce fragmentation in sustainability-related financial disclosure standards was indeed urgent. Our standards are based on what capital markets participants tell us is important and feasible through our public consultation processes.

Eccles: Did you hear anything from investors, given you are doing all the work for them?

Lloyd: Investor perspectives are essential to our process. During the trustees’ consultation we engaged heavily with them. We had more than 50 investment houses and investor associations commenting on the consultation. Almost 30% of comment letters received were from market participants.

Eccles: How is this investor support translated in practice?

Lloyd: In several forms. For example, some asset owners are encouraging their portfolio companies to disclose in accordance with ISSB Standards. Many investors engage in jurisdictional consultations about the adoption of ISSB Standards, including by sending comment letters noting the benefits of aligning local requirements with ISSB Standards. Some global investors refer to the ISSB in their investment guidelines and voting policies.

Eccles: Right, I have seen BlackRock policies calling for standardised disclosures.

BlackRock “Standardized disclosure of sustainability-related data supports investors in making informed decisions. The International Sustainability Standards Board (ISSB) standards, IFRS S1 and IFRS S2, represent one such approach to standardization that we find useful in our analysis”.

Source: BlackRock Investment Stewardship Global Principles for Benchmark Policies Effective as of January 2026. https://www.blackrock.com/corporate/literature/publication/blackrock-investment-stewardship-benchmark-global-principles.pdf

Eccles: However, you are not consolidating all the sustainability disclosure landscape. The most popular term since the creation of the ISSB is probably ‘interoperability’. I heard you saying several times that it is because the European Commission started to develop European Sustainability Reporting Standards (ESRS) before the ISSB was created. When the ISSB work started, the U.S. SEC also had its own work to develop a climate rule. And then there’s the Global Reporting Initiative (GRI), which takes a multi-stakeholder approach. Why are you investing so much time working on interoperability between ISSB Standards and other initiatives?

Lloyd: To ensure that investors receive comparable information and to limit the burden and costs for companies. As sustainability-related disclosure requirements come into effect around the world, reducing duplication and making requirements work together more effectively can significantly reduce the reporting burden for companies while providing high-quality information for investors. ISSB Standards provide the foundations for globally applicable disclosure requirements for investors, thereby establishing a consistent global baseline. Jurisdictions are then able to require disclosure of further information, by adding their own additional ‘building blocks’ to these global foundations, tailored to the incremental information needs that a jurisdiction considers necessary. Companies can also build on top of these disclosures to communicate to additional stakeholders, for example using the GRI Standards.

Eccles: Why did you use a building block approach when designing ISSB Standards?

Lloyd: Taking this building block approach provides an efficient reporting system for the benefit of companies, retaining alignment with ISSB Standards. That is the case for ESRS, for example. When the ISSB was established, we knew that our investor focused standards needed to work within this context – where parties other than investors may also be interested in sustainability-related information. That is why we are clear in our general standard (IFRS S1) that a company can still comply with ISSB Standards, as long as any incremental information that is beyond that for investors, is presented in a way that enables the globally comparable, investor-focused information to be clearly identified.

Eccles: You talk about alignment with ESRS despite a different approach to what companies have to report? In other words, despite a different approach to materiality?

Lloyd: That is right. This is a prime example of the building block approach. There are many disclosures common to our sets of standards as there are disclosures relevant to investors (our focus) and for the broader focus of ESRS (which covers both investor needs and other stakeholder information needs). ISSB Standards and ESRS use different overall materiality concepts, but they are substantially aligned when it comes to financial materiality and information relevant to investors.

Eccles: To what extent can companies reuse disclosures across these two different sets of standards?

Lloyd: We have focused on ensuring that for that common set of information that is financially material there is as much alignment as possible, meaning that companies reporting under one set of standards should generally be able to identify the disclosures required by the other, particularly for climate-related reporting. While ISSB Standards focus on the provision of information relevant to investors, or in other words, information that could influence decisions to invest in a company, ESRS also consider the provision of information regarding a company’s impacts on people and the environment that do not create risks or opportunities for a company. So, this means there is additional information required by ESRS that we do not require.

Eccles: So, do ISSB Standards ignore impacts?

Lloyd: No, they don’t. ISSB Standards acknowledge that a company’s risks and opportunities arise from its impacts and dependencies on resources and relationships throughout its value chain. Information about impacts and dependencies could be material to investors in understanding the sustainability-related risks and opportunities that could reasonably be expected to affect a company’s prospects.

Eccles: How does this work in practice? Is there a specific example you’re thinking of?

Lloyd: In the example of the company using water to operate its business, the company’s activities could impact a local marine species. If the company’s activities have the potential to contribute to the degradation or depletion of that water source and therefore endanger the marine species, these potential impacts could create a risk to the company’s operations because of its dependency on its legal and social licence to operate. So, the company might determine that these impacts give rise to a sustainability-related risk that could reasonably be expected to affect the amount and timing of the company’s incoming and outgoing cash flows, its ability to obtain funding and/or the cost of capital and, therefore, the company’s prospects. Of course, when companies are using other standards to meet information needs beyond the common information needs of investors, they might report on the impact on the marine species regardless of whether it creates a risk for the business or not.

Eccles: How does this compare to the requirements in ESRS?

Lloyd: What makes ISSB Standards different from ESRS is that reporting under the ISSB Standards is focused solely on providing material information about impacts if they give rise to sustainability-related risks and opportunities that could reasonably be expected to affect a company’s prospects (that is, its cash flows, cost of capital and access to finance) and when that information could reasonably be expected to influence investment decisions, whereas ESRS do not limit the disclosure of information about impacts in this way. They cover information on impacts more broadly including as an effort to support Europe’s policy objectives.

Eccles: In short, how does the philosophy of ISSB Standards differ from that of ESRS?

Lloyd: ISSB Standards were developed to provide sustainability-related financial disclosures that provide critical information for capital markets, whereas ESRS were also developed to achieve policy objectives which extend beyond providing information to capital markets. They can coexist efficiently.

Eccles: So, information about impacts might be relevant for investors. Of course there are lots of different types of investors. Some, including those with a mandate beyond financial return alone, might be interested in information not provided by companies applying ISSB Standards.

Lloyd: It is important to consider what financial reporting can do and what it cannot do. The ISSB’s focus, like the IASB’s focus, is to develop standards that - when applied by companies - provide investors with financially material decision-useful information so that they can make informed investment and capital allocation decisions. A company should provide information directly relevant to its business; that is, information that relates to the company’s prospects for economic success, without inherent bias towards any particular investment or capital allocation approach.

Eccles: When looking at groups outside the European Union doing business within the European Union, the next chapter in the ESRS sequence is ESRS-40a. Are the same considerations about interoperability between ISSB Standards and ESRS also valid between ISSB Standards and ESRS-40a?

Lloyd: As there are non-EU groups within the scope of ESRS-40a that already use ISSB Standards, or ISSB-aligned jurisdictional standards, we are engaging with EFRAG to identify efficient ways for these non-EU groups to use an “ISSB+” approach; that is, to enable them to start from their reports prepared applying ISSB Standards to add the information needed to comply with ESRS-40a.

Eccles: Is there any overlap?

Lloyd: Yes. Although the two sets of standards have different perspectives, with ISSB Standards focusing on sustainability-related risks and opportunities that could affect a company’s prospects and ESRS-40a focusing on impacts on people and the environment, there is overlap. EFRAG’s proposals for ESRS-40a identify investors as a user group for the information, alongside other stakeholders. And, as I talked about earlier, information about impacts can be relevant to explain sustainability-related risks and opportunities. As a result, there is an overlap in disclosures, particularly for climate-related matters.

Eccles: To conclude, can you tell me what good will look like in five years?

Lloyd: I hope that in five years, we won’t even need to have a debate about whether information about sustainability matters to capital markets. Having the ISSB’s global baseline used across major capital markets worldwide, delivering consistent, comparable data that investors rely on just as routinely as traditional financial statements is what good looks like to me.

Eccles: And how do you expect companies and investors will be using the disclosures?

Lloyd: We’ll see sustainability-related disclosure reported side-by-side with financial statements that is robust such that investors can rely on it. It can also be imagined that that enhanced quality of information will support strategic board discussions on company prospects that are already occurring. Ultimately, good is when transparent disclosure isn’t seen as a burden of compliance, but as the clearest path to understanding resilience, accountability, and long-term value creation.

Eccles: Which strategic ramifications do you see?

Lloyd: Investors in New York, Hong Kong, Tokyo, London or São Paulo can look at a company’s sustainability disclosure and evaluate it with the same trust and rigour as traditional financial metrics. It means we’ve built the global baseline we set out to build, and that it’s actively helping companies and investors navigate an uncertain world with confidence.

Eccles: This means, you hope sustainability reporting is no longer treated as a separate, extra exercise. It is simply business as usual. You mentioned robust disclosure, which provides investors with the confidence in the information disclosed that they ultimately use to make their investment decisions. What role does the ISSB play with assurance?

Lloyd: Our mandate does not include requiring a particular level of assurance or developing auditing/assurance standards. The question of whether information has to be assured and the level of assurance is determined by regulatory or jurisdictional authorities. This is also the case by the way for financial statements. The IASB does not decide on assurance there either. Our role is to set standards that enable the information to be assurable.

Eccles: What do you mean by ‘assurable’?

Lloyd: For example, a company is only required to disclose information about sustainability-related risks and opportunities identified by the company using reasonable and supportable information that is available at the reporting date without undue cost or effort. This was included in finalising the ISSB Standards to clearly specify the basis for identifying which sustainability-related risks and opportunities need to be reported on.

Eccles: We’ve covered a lot of ground in this conversation. Thanks again for your time, Sue. It was great catching up.

Lloyd: Thanks to you, particularly for the provocative questions. I enjoyed the conversation.

Portrait of Luca Pacioli, attributed to Jacopo de' Barbari, 1495. Museo di Capodimonte, Naples. Via Wikimedia Commons.

I recently read The Carbon Accounting Landscape Review, published by the International Chamber of Commerce (ICC) and prepared by the Secretariat of the Technical Expert Panel on Carbon Accounting, which ICC co-leads with Carbon Measures. It is a detailed and careful account of the current state of product-level carbon accounting, a field roughly 15 years old.

For whatever reason it reminded me of three pieces I wrote about four years ago with my colleague Kazbi Soonawalla at Saïd Business School, University of Oxford, published in the Harvard Law School Forum on Corporate Governance: The Long and Winding Road to Financial Reporting Standards, The Complex, Contentious, and Changing Nature of Financial Reporting Standards, and Putting Financial Reporting Standards Into Practical Perspective. These pieces set the formation of the International Sustainability Standards Board (ISSB) against the history of financial reporting:I went back and re-read them. Doing so took me back more than 800 years. What follows uses that lens.

This is not an assessment of the six approaches the Landscape Review examines, or of its technical conclusions. It is an assessment of what the history of standard setting suggests about the situation the report describes.

What the Landscape Review Says

The Panel was appointed jointly by ICC and Carbon Measures and is co-chaired by Amy Brachio and Karthik Ramanna. The report is careful about its own standing: a foundational resource for the Panel’s future deliberations rather than the Panel’s conclusions, a qualitative documentary analysis rather than quantitative testing, and a point-in-time assessment of a field it expects to keep moving.

It states its premise openly. The initiative behind it was founded on the conviction that ledger discipline is what product-level carbon accounting requires. The review therefore does not ask whether existing frameworks are any good on their own terms; it says plainly that they were built for other purposes and succeed at them. It asks what each contributes toward a ledger-based architecture, and what remains open.

The problem it identifies is framed in accounting rather than measurement terms, and the framing is the report’s rather than mine. A product carbon footprint, it argues, is not a directly measurable property of a product. Unlike sulfur content, you cannot test for it. It is “an informed accounting calculation” built from activity data, emission factors, supplier information, boundary choices, and allocation rules. So a procurement officer looking at two figures for two shipments of steel cannot tell from the product whether the difference reflects better carbon performance or different calculation choices. That is a contestable way to describe a life-cycle assessment, and a practitioner might answer that a footprint is a modelled physical quantity with uncertainty bounds rather than a construct. But it is the premise everything else rests on, and the report’s shorthand for what is missing follows from it: the equivalent of U.S. GAAP or IFRS for carbon.

The review examines six approaches: (1) ISO 14067, (2) the GHG Protocol Product Standard, (3) PACT v3, (4) the EU’s Product Environmental Footprint, (5) ISCC’s Carbon Footprint Certification, and (6) the E-Ledgers proto-standards with related work from the EFI Foundation and the Carbon Accounting Standards Initiative (CASI). It finds that the landscape is a strong starting point rather than a blank slate including mature life-cycle science, a shared allocation hierarchy, and working data-exchange infrastructure. It finds two things missing. The first is quantification—comparability strong enough to hold up between competing producers, a tightened allocation rulebook, and consistent treatment of the hard cases such as removals, biogenic carbon, capital goods, and circular flows. The second is architecture—formal transfer of embedded emissions between entities, an unresolved choice among chain-of-custody models, assurance at the level financial accounts are held to, and operability at the level of the transaction.

Two features of the report matter for the historical comparison. It is markedly more conciliatory than the surrounding public argument about competition between the two organizations and the motives attributed to them, especially Carbon Measures. The report acknowledges that ledger accounting is not the answer to everything the disclosure ecosystem addresses and, in setting the GHG Protocol Corporate Standard aside from the product-level assessment, that the Corporate Standard is the accepted foundation of entity-level climate reporting and that its value there is not in question. It is also candid about its own favored candidate. The approach it identifies as supplying the missing architecture is also the one it calls least proven, needing allocation and boundary rules, ledger specifications for enterprise systems, assurance protocols, and testing at scale. Table 4 from the report identifies what each approach contributes and what remains open.

The Standard-Setting Comparison

Those three pieces with Kazbi were written because the ISSB had just been established, EFRAG was drafting for the CSRD, and many people were asking whether a confusion of NGOs writing sustainability standards was simply being replaced by a confusion of government-backed bodies writing the same ones. The answer we gave was to set the preceding two years against a hundred and fifty years of financial reporting.

One distinction has to be made first. Entity-level carbon reporting and product-level carbon accounting are not the same enterprise at different magnifications. Their primary users differ. Entity-level disclosure exists for investors and regulators, and the GHG Protocol Corporate Standard has spent a quarter of a century becoming the accepted foundation for that. Product-level data is for a buyer choosing between two shipments of steel and a customs authority pricing an import. It is a continuum rather than a wall—entity-level Scope 3 figures are built from product data aggregating up the chain. But the institutions that confer authority at the two levels are not the same, and the GHG Protocol Product Standard dates from 2011. So does very nearly everything else in this field.

Six Lessons about Standard Setting

The features now being treated as peculiar to carbon (e.g., multiple bodies, methodological disagreement, and disputes over standing) recur throughout the history of standard setting. What follows is six things the history shows.

Lesson 1: Standards are Social Constructs

Kazbi and I put it this way in 2022, when the subject was the ISSB rather than carbon: “standards are social constructs, not something grounded in the laws of physics.” There may be technically worse answers. There is never a single scientifically correct one.

That determines what kind of disagreement this is. If a standard were a discoverable fact, two bodies working the same subject would be a contest, one would be right, and evidence would settle it. Standards are not discoverable facts. They are agreements about how to represent something, chosen from defensible alternatives, for a purpose. Two bodies working the same subject are therefore not a race but two attempts at an agreement.

Lesson 2: Product-level Carbon Accounting Has No Inherited Conceptual Base

Luca Pacioli published his Summa de arithmetica, geometria, proportioni et proportionalità in 1494, a compendium of Renaissance mathematics whose section on bookkeeping, Particularis de Computis et Scripturis, is the first printed description of double entry bookkeeping. He is routinely called the father of double-entry bookkeeping.

That title is misleading if taken to mean that he invented the practice. Italian double-entry methods are conventionally dated to at least 1211, and the direction of scholarship has been to push the date earlier rather than later. Genoa’s public accounts of 1340 were long treated as the earliest undisputed example, until a Florentine ledger from the Champagne fairs of 1296 and the Farolfi company’s books of 1299–1300 displaced them. A recent analysis in the Economic History Review argues that the 1211 ledger of Florentine moneychanger-bankers is itself in double entry. On any of these datings the practice preceded Pacioli by centuries.

It was six centuries after 1211 before professional accountancy bodies existed at all, first the Society of Accountants in Edinburgh in 1854 and followed by the Institute of Chartered Accountants in England and Wales in 1880. By 1926 more than 90 percent of industrial companies listed on the New York Stock Exchange were audited, though no law required it and the Exchange itself did not require audited statements from newly listed companies until 1933. As the accounting historian Stephen Zeff has documented, there was at the time no authoritative body of accounting standards to audit against. Each company chose its methods and each auditor applied its own. Nor was disclosure complete. Large listed firms including Westinghouse, Singer, and National Biscuit published no balance sheet, no income statement, or neither, and many withheld sales figures on the argument that competitors would use them.

Then came the crash of 1929, the Securities Acts of 1933 and 1934, the creation of the Securities and Exchange Commission, and only then, in 1939 after the regulatory demand existed rather than before it, the profession’s first standard-setting body: the Committee on Accounting Procedure. It issued 51 Accounting Research Bulletins before being replaced by the Accounting Principles Board 20 years later and, in 1973, by the Financial Accounting Standards Board (FASB). That same year, professional bodies from nine countries founded the International Accounting Standards Committee (IASC), which became the International Accounting Standards Board (IASB) in 2001. Four hundred and forty years separate Pacioli from the SEC. Across that history practice and institutions preceded any formal conceptual framework, and the standard setting that followed worked from an inherited bookkeeping architecture rather than devising one from scratch.

Product-level carbon accounting has no such inheritance. There is no equivalent of double entry to build on—no settled architecture for recording an emission once, attaching it to a product, and carrying it between the firms that handle that product; nothing in the six approaches the report examines yet supplies one. It is being asked to settle the conceptual base and write the standard on top of it at the same time, with CBAM’s definitive regime already running. The episodes described here offer no close precedent for settling both questions at once under an active regulatory deadline. The Landscape Review says as much about itself; its whole premise is that the accounting architecture is the missing piece.

There are candidates for that base, and at least one claims the double-entry lineage directly. That claim is not assessed here. The inheritance supplies bookkeeping mechanics; it does not by itself settle the allocation logic on which the six approaches the report examines still differ.

The diagram below shows product-level carbon accounting in historical context to financial accounting and reporting.

Two roads, drawn to the same scale. Financial reporting milestones from Eccles and Soonawalla (2022); carbon milestones from publication dates recorded in The Carbon Accounting Landscape Review, Section 5.1.

Lesson 3: The Hard Parts Are Hard because Interests Differ

Some disagreement reflects immaturity where the field is young and the data imperfect. Other disagreements persist because constituencies prefer different outcomes, and technical progress does not remove that conflict.

The cleanest illustration is other comprehensive income. Ninety years into mandatory financial reporting there is still no settled principle governing which gains and losses bypass the income statement, and recycling rules differ between IFRS and U.S. GAAP. This is not an obscure corner; it determines reported earnings. The FASB and the IASB worked on the conceptual framework that was supposed to resolve it and did not, because what belongs in earnings is a question about what earnings are for, and preparers, investors, auditors and regulators do not agree. Carbon has its own versions, which The Landscape Review names: allocation across co-products, the treatment of removals and biogenic carbon, and the choice among chain-of-custody models. Better data alone will not resolve those choices. As Kazbi and I put it, it is worth being “realistic about the effort, difficulties, and time it will take.”

Lesson 4: Standards Never Stop Changing

Major standards are routinely revised as transactions, practices, and policy needs change. A standard that never changed would eventually cease to fit what it was intended to describe. The accounting standards at issue here also do not themselves determine the policy targets to which the resulting measures may be applied. Whatever either body publishes will not be the last word, and treating a first standard as a final settlement misunderstands what a standard is.

Lesson 5: Companies Use Standards and then Report Around Them

In the third of our 2022 pieces we looked at what companies actually do with a mandated standard, and the answer is that they supplement it. Govindarajan, Srivastava and Zhao put the figure at roughly 95 percent of S&P 500 companies reporting non-GAAP earnings alongside their GAAP ones. The SEC’s response was not to forbid the practice but to govern it. Regulation G bars materially misleading non-GAAP presentations and requires reconciliation to the most directly comparable GAAP measure. A mandatory standard did not end the argument about measurement. It gave the argument a common reference point.

Lesson 6: Convergence Is Desirable, Hard, and Usually Ends in Interoperability

The FASB and the IASB signed the Norwalk Agreement in 2002, committing to converge U.S. GAAP and IFRS. The response was not unanimously positive even then. Critics worried about a one-size-fits-all approach and about the very different starting points of prospective adopters, and there was real apprehension that harmonization, a word denoting flexibility and coming together, would be quietly replaced by standardization, a word denoting something rather more rigid.

In 2004 the boards added the conceptual framework to the convergence agenda on the assumption it would be relatively straightforward, being the most abstract and least commercially contested part of the job. It proved the reverse. By 2010, on completing the first phase, both boards published their own conceptual frameworks and gave up on a common one. Some of the concrete projects fared better than the abstract one. Revenue recognition converged and has held. Financial instruments did not; leases produced two standards issued in the same year that still differ in structure; impairment produced expected credit losses on one side of the Atlantic and current expected credit losses on the other.

The wind-down produced public recrimination. In June 2014 the former SEC Chairman Christopher Cox blamed the IASB for the divergence, and the IASB chair Hans Hoogervorst responded that convergence had always been a limited-scope project with successes on revenue recognition and challenges on financial instruments.

In this history, success has meant not one text but texts close enough that the differences can be bridged. That is also where the present positions sit. The Landscape Review says its architecture is designed to be interoperable with existing systems rather than to replace them. Asked about fragmentation in an interview I did with Amy Brachio, she said she believes it can be addressed through a focus on interoperability. The GHG Protocol and ISO are going further still, consolidating two corporate standards into one. The disagreement is about how, not whether.

One factual difference between then and now is worth noting without drawing a moral from it. Norwalk was attempted after both bodies of standards had hardened. Here, neither product standard exists yet. The GHG Protocol and ISO have set an integrated public consultation on their consolidated corporate standard for the second quarter of 2027 and publication in the fourth quarter of 2028. Their product-level joint working group convened in July. Carbon Measures, through the Technical Expert Panel, has delivered its landscape report on schedule, with a model standard to follow. Nobody sees a draft product standard from either body before 2027.

How Standards Actually Acquire Authority

Underneath all of this sits the question of who is entitled to write a standard. That is the wrong question, because the historical answer to it is nobody in particular. The better question is how a written standard acquires authority, and there are at least six routes, all of which have worked. There is also an external benchmark for the process. In 2000 the WTO’s Committee on Technical Barriers to Trade adopted six principles for developing international standards—transparency, openness, impartiality and consensus, effectiveness and relevance, coherence, and the development dimension. Written for trade rather than carbon, they provide a benchmark for comparing standard-setting processes.

The first route is delegated statutory authority. A regulator keeps the legal power and hands over the drafting. The SEC delegates to the FASB, which is funded by mandatory accounting support fees rather than voluntary contributions from the companies it regulates or the accounting firms who do the audits. This is the entire point, because the two predecessor bodies, the Committee on Accounting Procedure and the Accounting Principles Board, were each displaced amid persistent doubts about their independence from the professional body that sponsored them. In this lineage, perceived sponsor dependence was sufficient to undermine institutional legitimacy.

The second is endorsement and jurisdictional adoption. Where no single regulator has jurisdiction, an international regulators’ body endorses and governments adopt. IOSCO endorsed the international standards in 2000; the EU mandated them in 2002 with effect from 2005, They are now required in more than 140 jurisdictions. That is roughly 30 years from the founding of the IASC to the point where the standards had force. And legal force came through regulators’ endorsement and legislative adoption rather than from the standard setter acting alone.

The third is de facto adoption through the ecosystem, and it involves no mandate at all. The GHG Protocol was founded in 1998 by an NGO and a business membership organization, explicitly to head off a proliferation of incompatible corporate inventory methods. No government created it and none has ever been required to approve it. Its authority grew as CDP, then the SBTi, then regulators and eventually the drafters of the CSRD pointed at it, and the ISSB wrote it into IFRS S2 as the required basis for measuring emissions. Its formal governance arrived afterwards and is now substantial: a Steering Committee that appoints and oversees an Independent Standards Board, Technical Working Groups, a written development and revision procedure, and public consultations weighed against published criteria rather than counted as votes. Legitimacy through use first, procedural formalization second.

The fourth is less familiar outside the technology world. Here an industry consortium writes a specification, the market adopts it, and a formal standards body transposes it into an international standard. ISO/IEC’s Joint Technical Committee 1 (JTC1) runs a Publicly Available Specification process for exactly this. Note the order. JTC1 first vets the submitting body against openness criteria, and only then accepts its specifications which enter at Draft International Standard stage for national ballot. More than 115 ISO/IEC standards have come this way. It is not a rubber stamp, and it is not immune to capture either. Office Open XML, approved in 2006 by Ecma International, the industry consortium behind the JavaScript and C# specifications, and pushed through the JTC1 fast track in 2007 and 2008, produced one of the most contested ballots in the organization’s history, with formal appeals lodged by four national bodies. The safeguard in this route is the transposing body’s own due process rather than reliance on the consortium’s.

The fifth is codification of established practice, which is where this piece started. Use comes first, someone writes it down, and authority follows from the demonstrated fact that it worked. Pacioli.

The sixth is the one operating right now, which is why it is odd that it features so little in the argument. A regulator writes its own methodology and waits for nobody. CBAM does not adopt ISO 14067 or the GHG Protocol Product Standard; it prescribes its own. California’s Low Carbon Fuel Standard runs on CA-GREET4.0. China builds on ISO 14067 through a national standard and its own emission-factor database. The Landscape Review’s explanation is worth taking seriously; no existing standard has been consistently referenced to underpin product-level compliance, so policymakers are defining bespoke approaches of their own. Its conclusion is that what needs harmonizing is the data layer underneath the regimes rather than the regulatory designs on top of it.

Routes three, four and five share one feature: use precedes or contributes to formal authority. They remain distinct in what follows—ecosystem dependence in the third, transposition by a standards body in the fourth, and codification of established practice in the fifth. The GHG Protocol’s own history is the nearest example of the first, and the proposition that a standard is only legitimate if a government commissioned it does not survive contact with the record. Existing use is itself one source of authority, and at entity level the GHG Protocol has long since established it. That is not what product-level accounting is about.

Neither body has yet published the process by which this particular standard will be developed, consulted on, and approved. The GHG Protocol has a written development and revision procedure, a published governance structure, and, since July, a published development plan for the consolidated corporate standard. It has published no equivalent for the product-level work, which is run jointly with ISO under a partnership less than a year old, and how that joint process will handle technical disagreement is not yet public. Carbon Measures has a Panel whose composition and mandate are public and whose first report is out, but the process for developing a model standard has not been described. Without those process descriptions, judgments about comparative procedural legitimacy remain premature.

What Makes This Case Different from the Analogies

The principal institutional comparisons used so far involve bodies built for continuing roles. That is not the situation here. Carbon Measures has said publicly that it expects to exist for five to seven years and then close. In my interview with Amy Brachio she described success as a framework adopted globally and carbon intensity standards in place in multiple markets; the organization’s mission is policy, and product-level accounting is the instrument. When I interviewed Karthik Ramanna he said the E-Ledgers Institute would be delighted to dissolve itself once competitive differentiation on carbon becomes mainstream.

That changes the shape of the problem. One of the two organizations is not bidding to become the permanent standard setter and has put a clock on itself. The other is a permanent institution which, by its constitution, cannot advocate for the product regulation, border measures, or intensity standards that would give product-level accounting its bite. It also clarifies who Carbon Measures is addressing, which is not the audience most commentary assumes: not investors, and not in the first instance companies, but the bodies capable of setting carbon intensity standards. Brachio has been explicit that Carbon Measures advocates for governments to regulate using product intensity standards, and that each country would set its own on the products that matter to its industries.

It also raises a design question that is largely absent from the current argument. If an organization builds an accounting architecture and then dissolves, who maintains it? Standards require continuous revision, as everything above suggests, and somebody has to do the revising. A self-liquidating body therefore has to solve not only adoption but succession, and no public account of how that would work yet exists.

If two materially different product-level carbon standards emerge and stay different, the cost will be borne by procurement officers trying to compare quotations, customs authorities trying to verify declarations, and producers running parallel accounting systems to sell into different markets. Those are the parties that bear the operational cost of persistent divergence.

What Happens Next?

Here are three things I don’t know. I don’t know the extent to which the GHG Protocol, ISO, and Carbon Measures will collaborate with each other in developing a product-level standard for carbon accounting. I don’t know how similar or different the draft standards they issue will be. I don’t know which route each set of standards will be taking for gaining social legitimacy.

Here are three things I do know. I know that the GHG Protocol and ISO are standard setting organizations and that Carbon Measures is not. I know that each standard setting process will be difficult and contentious. I know that if we get to one global set of standards, which I hope we will, they will not be locked down but will continue to evolve for years and decades to come.

This is a guest contribution by Beate van Loo-Born, Founder of BvLB Advisory and former UBS executive

I have been working on this since before it had a name. My first thesis, on eco-industrial development, was written in 2003, a year before the UN Global Compact’s Who Cares Wins report supplied the acronym ESG. Industrial ecology asks how to design a system so that one actor’s waste becomes another’s input. It is about closing loops and aligning incentives, which is what sustainability ought to mean, rather than the art of persuading people to behave better.

Twenty-three years on, the field has far more vocabulary but still very little feedback. I spent most of those decades in executive roles in financial services, at UBS and then at the Swiss stock exchange, before becoming chief operating officer of a philanthropically funded climate data organisation backed by two large climate foundations. I did and do believe in the mission. But we were flying blind in a dysfunctional system.

I moved from there into manufacturing. In finance and in manufacturing alike I never had to wonder whether a product was useful. Customers renewed or they left, revenue held or it fell. Where markets move slowly and products last decades that verdict takes years rather than quarters, but you receive the feedback eventually.

Climate philanthropy is spending six billion dollars a year on theories largely untested, given to organisations that do not close when they fail. The sector has yet to build the machinery that markets in the real economy use for recalibration and consolidation as a matter of course. The system is dysfunctional, which is a design failure rather than a moral dilemma. There are five failures, set out below, and five remedies that are already well established in other sectors.

Five Failures

1. User-buyer split

In a market the person who uses a product also pays for it, which is how a user’s opinion acquires force over a producer. When the thing stops being useful the money stops, and the producer has to find out why.

Philanthropy splits the two. The organisation I worked for produced analysis of corporate climate performance meant to inform lending and investment decisions. Its intended users were banks and investors. Its payers were foundations. A great many organisations in this sector are built to that shape, and none charges the people it is designed to serve.

Clients of these organisations often do not come looking for the product, because it exists to change how they behave rather than to help them get better at what they are already doing. That makes willingness to pay a weak test of whether it works.

Take the case above. An investor is trying to earn a return, not to cut emissions. Offered climate data for free they will take it, but impact is only created if they act on it, and an investment process draws on a great many data points, most of them there to optimise for alpha.

So the funder learns neither that the product is any good nor that the target has been reached. Climate impact is also a shared target of the entire sector, so any single organisation’s effect on the whole problem is diluted further still.

Climate is what Rittel and Webber called a wicked problem in 1973: no definitive formulation, no stopping rule, no clear owner, and no immediate or ultimate test of a solution. That last property has become the sector’s standing defence. If the problem admits no final verdict, the argument runs, asking for evidence of impact is naive.

It is a real difficulty and a poor excuse. The absence of a final verdict does not rule out an interim one. An investor either uses the data or does not. A bill either gets introduced or it does not. Wickedness explains why this sector cannot measure its ultimate effect on the climate. It does not explain why it declines to measure the steps in between.

2. The No-Exit-Curse

If you are not running a business, you cannot go out of business. A firm that cannot fund itself runs out of money, and its people, data and equipment are released back into the economy. Some see this as an unfortunate by-product of capitalism, but it is most of what makes a market economy intelligent, and philanthropy has no substitute for it.

Roughly half of new American businesses are gone within five years, and every one of those closures frees people and resources for a more productive venture. Nonprofits close too, but they close when a funder changes its mind, which is a verdict on a relationship rather than on results.

Nothing in this sector dies on the grounds of not working. I went looking for a counterexample and could not find a climate organisation that closed saying its theory of change had been tested and had failed. Plenty closed for other reasons. The Net Zero Banking Alliance ceased operations in October 2025 after its members left under political pressure. Climate Generation closed in February 2026, describing twenty years of meaningful impact on the way out.

Nor does hardship force consolidation. Bridgespan found nonprofit merger rates essentially unchanged between 2001 to 2006 and 2007 to 2012, because filings rose but registrations rose faster. The sector consolidated a little and proliferated more, through the worst financial crisis in seventy years.

Meanwhile money keeps entering and more organisations are founded each year. Bob counts six to seven thousand registered environmental NGOs in the United States alone, against six billion dollars of foundation climate funding worldwide in 2024. A global numerator against an American denominator gives roughly a million dollars each, and the real figure is lower. Funding has more than doubled since 2020 without producing fewer and larger organisations.

3. Competition without a price

Competition works when competitors are chasing the same person’s money and that person is the one being served, because the price then carries the answer back to everyone at once. Where the money comes from a funder there is no price at all, so working harder to win means working harder to impress the funder, which is not the same activity as working harder to be useful. Two organisations building substantially the same product are not a market discovering the better one.

Redundancy is genuine insurance, but only when somebody chose it. Nobody surveyed this field, concluded that four competing corporate assessment providers was the right number, and funded four. Unselected redundancy is not resilience but cost.

4. A conglomerate with no head office

Coordination took me longest to understand. I assumed that organisations sharing a cause would behave as collaborators who occasionally competed. Quite the contrary. There are countless collaboration meetings across the sector, spanning the traveling circus of COP and Climate Week, and very little real commitment or accountability to deliver anything together. How could there be, when nothing enforces either between organisations competing for the same funder?

Unlike a conglomerate, philanthropy has no head office. Each funder sees only its own portfolio, and the people who can see the duplication are the grantees sitting in those meetings, who have no standing to name it, because naming it means arguing that a neighbouring organisation should be defunded.

5. No theory of when it is wrong

The best objection is that philanthropy exists precisely to fund what has no price. Advocacy, litigation and movement building have no paying customer and no verdict for years, and they produced the sector’s biggest wins. Market discipline would have killed Beyond Coal in year three, and by 2020 that campaign had retired 318 of 530 US coal plants, double the goal set in 2011.

But nobody can say how much of that retirement Beyond Coal actually caused. Coal collapsed during the shale boom, and the two causes are not even independent, because the Sierra Club received more than 26 million dollars from sources connected to Chesapeake Energy while promoting gas as the replacement for coal. If attribution cannot be settled for the best-documented result the sector has, it can hardly be settled anywhere else.

The answer is not to measure everything. Unpriced does not mean unfalsifiable, and feedback is not the same thing as measurement. A litigation strategy has a docket, a movement has a membership count, an advocacy campaign has a bill number. What is missing is not metrics, of which there are already far too many, but any pre-committed statement of what would count as the theory having failed, and any actor with the standing to call it.

Why everyone is at COP - the veil of collaboration

One part of this system works, and it shows what the sector is capable of when a loop does close. Being seen by funders brings money, and money pays for the staff, the travel and the reports that get you seen again.

Senior people spend an extraordinary share of their time at climate conferences and the side events around them. Reading this as self-indulgence is mostly wrong. They go because visibility to funders is the closest thing this sector has to a sales channel, and because it is the only closed loop in the system. Energy flows to closed loops.

A company board sits on capital at risk, which gives it a selfish reason to want bad news early. In a philanthropically funded organisation the only stakeholder with real power is the funder, whose reputation is invested in the grantee succeeding. The escalation path for anyone raising a concern terminates at the party least motivated to hear the answer. A good enough cause also excuses a great deal, because work judged poor anywhere else survives when criticising it looks like criticising the mission.

Five Established Remedies

None of these needs inventing. All five have a track record in other sectors, and the first costs nothing.

1. A falsification condition

The precedent is the U.S. Toxic Release Inventory of 1986, which required factories to publish their hazardous emissions annually. No law against the emissions, no fines, no determination of safe levels, just information, and emissions fell forty percent by 1990. The philanthropic version is one sentence, published in advance, for each grant above a threshold, stating what would tell the funder the theory was wrong. For a data product aimed at investors: if by year three nobody can name a single investment decision that changed because of it, the theory was wrong. It is written by the funder rather than the grantee, so the burden falls on the party that currently carries none.

2. Balance sheet finance

Land that matters ecologically comes up for sale on the seller’s timetable. The public agency or land trust that should end up owning it cannot move at that speed, because appropriations and fundraising take years. The Conservation Fund borrows cheaply and buys it immediately, holds it, then sells it on to that permanent owner once their money is in place, and the capital returns for the next purchase. Forty years, more than 4,000 transactions, never a year in the red.

A grant is spent once. Capital that has to come back is spent repeatedly, and it stops coming back when the deals stop working, which is a verdict a grant never delivers.

3. Usage of advance market commitments

An advance market commitment is a promise to buy something that does not yet exist, at a guaranteed price, so that suppliers have a reason to build the capacity to make it. The device was invented for vaccines, where guaranteed purchases persuaded manufacturers to build capacity for countries that could not otherwise pay.

Frontier applies it to carbon removal and has committed 1.8 billion dollars through 2040. What matters here is not the size of the commitment but that Frontier publishes what it has contracted against what has actually been delivered, on a schedule anyone can watch. Philanthropy has the same gap, between what a grant was funded to achieve and what it achieved, and never calculates it.

4. Specific consolidation

Consolidation covers three separate moves.

Function first. Three organisations each assessing a thousand companies by three different methods produce three sets of data that cannot be laid side by side. A study of six major ESG raters, memorably titled Aggregate Confusion, found their ratings correlated at about 0.54, and that more than half of the disagreement came from measuring the same attribute in different ways. One organisation assessing three thousand companies by one method is worth more than the three combined, because comparability is what is useful for transparency. The commercial side has made steps in the right direction: S&P bought Trucost in 2016, and in 2019 MSCI bought Carbon Delta and Moody’s bought Four Twenty Seven. The philanthropically funded side pools data through coalitions and benchmarks, and every provider survives rather than merging.

The standard setters have meanwhile conceded the point. The Climate Disclosure Standards Board and the Value Reporting Foundation were asking companies for much the same information in different shapes, and in 2022 both were folded into the IFRS Foundation. They recognised they had created mismatched demands and consolidated to remove them. Europe kept its own standard, then cut its mandatory datapoints by more than half in 2026. The organisations collecting the data have yet to follow.

Funding second, meaning fewer, larger and longer commitments. A two-year grant does not merely provide less money than a five-year grant, it changes what the organisation does with its time, because an organisation eighteen months from the end of its funding is one whose senior people are fundraising rather than working. Ford Foundation tested this with a billion dollars of five-year grants from 2016, and 83 percent of the organisations funded reported being more financially resilient at the end than at the start

Organisations last, because mergers are slow and expensive, and because a sector of fewer and larger organisations depends on the judgement of fewer funders. When they do happen they can work: a study of 25 Chicago nonprofit mergers found 88 percent of the merged organisations felt stronger afterwards. Word of caution from the real economy: Consolidation is not a virtue in itself. Where two organisations are doing the same job, merge them; where they are not, leave them alone.

5. Professionalise the funder rather than only the grantee

Every remedy above lands on the grantee. This one is for the people writing the cheques. A profession has shared definitions, published methods, entry standards and consequences for failure. Climate funders have fragments of the first two and nothing of the last two. ClimateWorks, six annual editions into tracking the field, puts 2024 mitigation giving somewhere between 11.7 and 18.4 billion dollars, and reports that it still cannot separate general operating support from unrestricted funding because funders label them inconsistently. An accounting standard makes revenue mean the same thing at two companies. Nothing does that for a grant.

Development aid fixed this in 2008. The International Aid Transparency Initiative gives donors one format for reporting what they funded, for what, and what came of it; more than 1,500 organisations use it. Climate philanthropy could adopt it tomorrow. Bob and Shivaram Rajgopal argued in June that corporate sustainability officers need to behave like a profession. It applies with more force to funders, because no one else in this system can change it.

Ordinary Practice in the Real World

None of these mechanisms are exotic. In the commercial and financial economy they are ordinary practice, which is what made the return to manufacturing so striking after four years without them. A company finds out whether its product works because someone either buys it or does not. Failure releases people and capital to something with a better claim on them. Boards want bad news early, because their own money is at risk. Revenue means the same thing at two different companies, because accounting standards make it so, and that is the reason their performance can be compared at all.

Industrial ecology, the discipline I started in twenty-three years ago, exists because linear systems, the ones that take materials in, use them once and discard them, eventually run down however well-intentioned their operators are. The fix is never exhortation but designing the system to close the loop.

Climate philanthropy is a linear system. Money enters, organisations accumulate, and nothing is consumed, nothing is recycled, nothing dies.

The sector argues endlessly about where the money should go and never about whether the last allocation worked. One sentence from the funder, written before the cash is allocated, saying what would show the theory was wrong, would give this sector its first routine way of finding out.

The Author

Beate van Loo-Born is the founder of BvLB Advisory. She spent twenty years in executive roles in financial services, at UBS and then at the Swiss stock exchange, before becoming chief operating officer of a philanthropically funded climate data organisation backed by two large climate foundations. She has since held executive roles in manufacturing.

She holds a MA in Sustainability Leadership from the University of Cambridge, where Robert Eccles was her thesis advisor, and wrote her first thesis, on eco-industrial development, in 2003 for an MBA at Clark. She holds an MA from the Sorbonne, an Executive MBA in Finance, and is conducting doctoral research on dysfunctional ecosystems at the University of Antwerp.

This is a guest contribution by Larry Selzer, President and CEO of The Conservation Fund

Twice before, America’s response to environmental challenges has changed course as circumstances have changed. Both times, the country adjusted and moved forward. We must do so again now.

The first transformation built our national capacity to conserve extraordinary places. The second created a legal and regulatory system capable of protecting people from pollution. Both were necessary. But the challenges now before us require another fundamental course correction—one that will drive the convergence of the environmental movement and the free enterprise system.

The First Era: Protecting Land From People

The first transformation occurred more than 120 years ago. In 1903, responding to the decimation by plume hunters of pelicans, herons, egrets and other bird species, Teddy Roosevelt established Pelican Island as the nation’s first federal bird reservation and the first piece of what became the National Wildlife Refuge System. Two years later, responding to the reckless destruction of American timberlands, he created the US Forest Service.

These actions kicked off the first Era of Conservation in America. It has remained a core national value ever since. Roosevelt recognized that a prosperous country could not consume its natural resources without restraint. He used the power of the federal government to establish over 100 national forests, wildlife refuges, parks and monuments.

This was a profound achievement. Roosevelt established that some places were too important to succumb to short-term economic interests. The national parks became the defining institutions of this era, grounded in the proposition that the country’s most magnificent landscapes should belong to everyone and be preserved for future generations.

The Second Era: Protecting People from Pollution

As the country industrialized, environmental harm came into sharp focus. Air pollution was making people sick. Industrial waste was contaminating rivers. Chemicals were moving through soil, water, crops, wildlife and human bodies. Economic growth had brought prosperity, but it had also imposed significant environmental and public health costs.

That gave rise to a second major shift.

The catalyst was Rachel Carson’s Silent Spring, published in 1962. Her seminal work gave Americans a new understanding of environmental risks. She connected environmental harm to human health and helped unleash a tidal wave of grassroots activism. That movement produced the first Earth Day and led to the passage of some of the country’s most important legislation, including the Clean Air Act, Clean Water Act, and Endangered Species Act. We now have tens of thousands of environmental laws, rules, regulations, and codes at the federal, state and local levels.

This second transformation achieved significant results. Air and water became cleaner, species recovered, and communities were better protected from environmental harm. This was the Era of Regulation.

The tools and institutions of this era were primarily designed to stop bad things from happening. For more than 50 years, the environmental movement has used the tools of legislation, litigation, and regulation to perfect the art of saying “No.” But these tools have now become the movement’s Achilles Heel. We have enormous capacity to oppose projects, but far less capacity to identify and advance projects that should proceed. That imbalance now threatens both the country’s progress and the environmental movement’s continued relevance.

The Third Era: Convergence

The country’s needs have changed again. The United States now faces demands that require us to build at a scale we have not seen in generations.

Addressing climate change requires new clean energy generation, transmission, and storage, as well as the electrification of our transportation system. Rising electricity demand from artificial intelligence, domestic manufacturing, and broader economic growth requires a larger and more reliable electricity grid. Supply chain vulnerability requires new factories and processing operations closer to home. Communities need housing, water systems, upgraded roads and bridges, and protection from extreme weather.

Every one of these has a physical footprint, and the scale is immense. Each requires capital, land, permits, infrastructure and public legitimacy. Each will affect communities and natural systems.

The electric grid illustrates the magnitude of the challenge. The Interstate Highway System, the country’s last large national infrastructure project, totals nearly 49,000 miles. Modernizing the electricity grid will require tens of thousands of miles of new high-voltage lines and upgrades to roughly 100,000 miles of existing lines over the coming years.

These new energy systems, transmission lines, manufacturing facilities, housing, and other infrastructure will affect forests, farms, rivers, and our other natural resources. Those impacts must be taken seriously.

But the answer cannot be simply to refuse to build. We cannot meet our climate goals without new energy infrastructure. We cannot drive economic growth without using land and other natural resources. We cannot fortify the resilience of our communities without investing in both natural and built infrastructure.

A movement organized around opposition, around saying “No” cannot lead the country though this transformation.

The country needs institutions capable of identifying not just what should be stopped, but what should be built, where it should be located, and how it can proceed while protecting essential natural, cultural, and community resources.

In short, we must learn how to say “Yes” without sacrificing our critical natural systems.

The challenge is not theoretical. At the end of 2025, approximately 8,200 power generation and storage projects were waiting to connect to the electric grid, representing more than 2,000 gigawatts of potential capacity. The queue demonstrates both the appetite to invest and the difficulty of translating investment into operating infrastructure.

The next era of the environmental movement, the Era of Convergence, must begin with a new proposition. Conservation must be integrated with the free enterprise system.

Convergence is the marrying of capital and conservation. It means bringing finance, markets, and free enterprise together with conservation. It also means bringing environmental organizations into the work of designing and enabling economic activity early, rather than engaging only after projects have been proposed.

The conserved working forest, restored wetland, and protected watershed are the physical manifestations of convergence. The convergence itself occurs in the way capital is aggregated and deployed. Climate and natural resource availability risks are valued. As transactions are structured, conservation is incorporated into economic decisions.

This is the work we do at The Conservation Fund. We are chartered for the dual purposes of environmental protection and economic development. We use private and public capital, land transactions, and market-based structures to move conservation from aspiration to execution. The model works because conservation is integrated in core economic activity. It is embedded in how land, capital, and risk are managed.

Convergence changes the role of conservation in development. Conservation can do more than mitigate damage after decisions have been made. When incorporated early, it can help inform where projects should go, identify places that should remain intact, protect the water and natural systems on which infrastructure depends, and create tangible benefits for affected communities.

Convergence does not mean abandoning environmental protections or accepting development everywhere. It means being as serious about advancing responsible projects in the right places as we are about opposing harmful projects in the wrong places.

Conservation, Capital, and Resilience

The insurance industry is beginning to understand that healthy natural systems reduce risk. Conversely, degraded forests, wetlands, flood plains, and watersheds will translate into larger physical losses. According to Swiss Re, the United States accounts for over 80 percent of global insured natural catastrophe losses, with approximately $89 billion in insured loss and $133 billion in total economic loss in 2025 alone. Insured natural catastrophe losses are projected to continue to grow by 5 to 7 percent annually in real, inflation-adjusted terms. Compounded, that implies a 28-40 percent increase after five years.

The National Institute of Building Sciences has reported that federally funded hazard-mitigation grants produce an average of approximately six dollars in benefits for every one dollar invested. Natural systems such as wetlands, forests, floodplains, and working lands form critical natural mitigation infrastructure that reduces flood, drought, wildfire, and water security risks and moderates damage when disaster occurs.

A manufacturing facility requires more than land and electricity. It needs reliable water, transportation, workers and infrastructure capable of withstanding disruption. A water utility depends not only on treatment plants and pipes, but also on the upstream watershed supplying its water. Transportation systems depend on wetlands and floodplains to bolster their resilience against storms.

The benefits can converge in a single project. In Tillamook, Oregon, the restoration of 443 acres of tidal wetlands improved salmon habitat, reduced flooding across 4,800 surrounding acres, helped protect a major highway and generated nearly $15 million in economic activity. Natural systems perform economically valuable functions that require investment alongside conventional infrastructure.

These natural systems are not amenities. They are necessities. When insurers, utilities, financial institutions, and corporations value those functions and direct capital toward protecting them, conservation moves from the margins of the economy into the machinery of economic decision-making. That is convergence.

Learning to Say “Yes”

The Era of Convergence will require a dramatic shift in how we approach environmental protection in this country.

The environmental movement must learn to support projects that enable growth, prosperity, and opportunity for local communities, while protecting critical natural, cultural and historic resources.

Learning to say “Yes” does not mean saying yes to everything. It means making choices. Some places are too important to develop. Some projects are poorly designed or improperly located. But we must become as capable of advancing responsible projects in appropriate places as we are at opposing harmful ones.

Businesses, investors, and governments must also change how they approach development. Businesses must engage communities early and honestly. Investors must better evaluate long-term ecological risk and resilience. Government agencies must establish clear and efficient pathways for responsible development. Communities must have a meaningful voice and a meaningful share in the benefits.

Convergence begins when environmental organizations and the institutions of free enterprise stop treating one another as opposing forces and begin working together to solve problems that neither can solve alone.

A New Conservation Movement

Teddy Roosevelt’s era taught us to protect the country’s natural inheritance.

Rachel Carson’s era taught us that environmental harm can dramatically impact human health, and reasonable rules are necessary to protect the public.

The Era of Convergence must give us the capacity to mobilize capital and conservation together to build the future of this country.

The Author

Larry Selzer leads The Conservation Fund with a simple conviction: protecting land and building strong communities are complementary goals.

As President and CEO, Selzer has helped The Conservation Fund protect more than 9 million acres of at-risk land valued at $8 billion across all 50 states. The organization’s work spans working forests and farms, recreation areas, wildlife habitat, historic landscapes, and urban green spaces.

Under his leadership, TCF has earned a reputation for speed, creativity, and impact, “finding the yes” in any complex conservation challenge. Selzer has expanded the organization’s reach to include the outdoor recreation economy, the clean energy transition, the protection of Native American cultural sites and African American historical sites, and initiatives that strengthen the connection between people and the outdoors.

Before joining The Conservation Fund, Selzer worked as a marine biologist at the Manomet Center for Conservation Sciences. He holds an MBA from the University of Virginia Darden School of Business and a Bachelor of Science in environmental studies from Wesleyan University. He serves on the board of Weyerhaeuser, is the board chairman of Leading Harvest and is an avid runner, hunter, fisherman and birder.

Chapters 8 through 11 were constructed as stress tests. Each scenario isolated a different configuration of political control and institutional response in order to observe how pressure moves when particular constraints are tightened or released. They are not forecasts, and they are not intended to suggest that one political outcome is more likely, more legitimate, or more desirable than another. Their purpose was simply diagnostic. By holding certain features constant and varying others, the scenarios made visible how the system behaves under strain.

What emerges from those exercises is not a contest among futures, but a set of recurring structural patterns. Across otherwise distinct political configurations, similar forms of delay, distortion, and reallocation appear. Pressure does not disappear when authority fragments, nor does it move cleanly or predictably when authority briefly consolidates. Instead, the system adapts to constraint. It reroutes influence, shifts the timing of effects, and redistributes responsibility across institutional boundaries. The scenarios differ in surface detail, but they converge in what they reveal about how the system absorbs and transmits pressure over time.

This chapter steps back from the scenarios to examine those patterns directly. Its focus is not on which configuration prevails, but on what divided government does to the operation of the system as a whole. Divided government is treated here not as a political outcome to be evaluated, but as a structural condition with identifiable effects on transmission, timing, and observability. It changes how influence travels, which edges become sensitive, and where pressure accumulates or dissipates. In that sense, it functions less as a blockage than as a modifier of system behavior.

This chapter begins with a longitudinal analysis of political structures since the establishment of the two-party system of Democrats and Republicans. The data show that divided or partially divided control is no longer an exception punctuating long periods of unified authority. It has become a dominant and persistent feature of contemporary governance. Authority is frequently fragmented across parties and often reverses before policies mature or feedback becomes legible. The result is a governing environment in which durable resolution is structurally difficult, even when political mandates appear momentarily strong.

The chapter then uses the Seven Edge System to provide an analytical lens through which this condition can be understood from the perspective of policies for climate change. Rather than asking whether divided government enables or constrains particular outcomes, the framework asks how fragmentation alters transmission across institutional boundaries. Which edges become more permeable? Which become sites of attenuation or delay? Where does pressure accelerate, and where does it dissipate into ambiguity? The answers to these questions do not lie in any single scenario. They lie in the patterns that recur across them.

What follows is therefore a synthesis, not a recap. The analysis that follows does not restate the scenarios or adjudicate among them. It extracts what they collectively reveal about system behavior under sustained fragmentation. By examining how divided government reshapes the operation of the seven edges, this chapter clarifies the structural context within which contemporary ESG conflict persists and the limits within which it can be resolved.

To ground this synthesis, it is first necessary to understand the structural condition at issue more precisely. The scenarios assume divided government as a given, but they do not explain how or why it has become so prevalent, nor what distinguishes contemporary fragmentation from earlier periods of mixed control. Before examining how divided government reshapes system behavior, the chapter therefore turns to the political science literature to establish its historical frequency, underlying causes, and conditions of persistence. What follows is not an interpretation of ESG conflict, but a baseline account of the governing environment in which that conflict now unfolds.

The History of Political Configurations

The distribution of federal governmental control configurations between 1857 and 2027 reveals a fundamental transformation in American political structure. Over 108 Congresses spanning six distinct eras, the balance between unified and divided government has shifted dramatically, with particularly striking changes emerging in recent decades. As shown in Table 3, across the full 170-year period, unified party control of the Presidency and both chambers of Congress occurred in forty eight Congresses (44% of the total), while divided configurations—whether split Congress (one party has the Presidency and either the House or the Senate) or opposition Congress (one party has the Presidency and the other both chambers)—characterized sixty Congresses (56%). Of these divided configurations, eighteen Congresses (16.7% of the total 108) were opposition Congresses. This aggregate balance, however, obscures profound temporal variation as shown in Table 4.

Era 1: Industrial Consolidation (1857-1911) demonstrates the most sustained pattern of unified control in the historical record. Unified government appeared in sixteen of twenty-seven Congresses (59%), with unified runs averaging eight and a half years in duration and the longest stretch extending fourteen years. Divided government configurations characterized 40% of the era (33% divided control, 7% opposition Congress), but these episodes were relatively brief and episodic rather than structurally entrenched. The balance slightly favored unified Republican control (ten Congresses) over unified Democratic control (six Congresses), reflecting the broader partisan dynamics of the post-Civil War period through the Progressive Era. This era established unified control as the normal baseline from which later periods would deviate.

Era 2: WWI & Interwar (1911-1933) represents a transitional period. Unified control appeared in five of eleven Congresses (45%), exactly balanced with divided control at 45%, and opposition Congress at 10%. Non-unified configurations thus characterized 54% of the era. The average duration of unified runs was six years—notably shorter than Era 1’s eight and half years but still substantially longer than what would emerge in later eras. This balance between unified and divided configurations foreshadowed the structural shifts that would become more pronounced in subsequent periods.

Era 3: New Deal & Post-War (1933-1947) stands alone in the historical record as the only era experiencing uninterrupted unified party control. All seven Congresses showed unified Democratic control, representing a continuous fourteen-year run that equals the longest unified stretch observed in Era 1 but, critically, occurred without interruption. This period witnessed zero instances of divided or opposition Congress configurations—a 0% divided government rate unmatched in any other era. The structural factors enabling this concentration—economic crisis, war mobilization, and partisan realignment—created conditions that proved unrepeatable in subsequent periods, even as unified control itself would return in later eras.

Era 4: Cold War Stability (1947-1979) marked the decisive inflection point toward divided government as a regular feature of American politics. Unified control appeared in only six of sixteen Congresses (38%), while divided government became the modal configuration at 50% of Congresses—the first era in which divided control exceeded unified control. Opposition Congress remained relatively rare at 13%. Non-unified configurations thus characterized 63% of the era, a substantial increase from Era 1’s 40%. The average duration of unified runs fell to four years—less than half the Era 1 average—though the longest run still reached eight years, indicating that sustained unified control remained possible even as it became less common. The balance between unified Democratic control (four Congresses) and unified Republican control (two Congresses) was less lopsided than in Era 1, reflecting increased partisan competitiveness.

Era 5: Post-Stagflation (1979-1993) represents the extreme point of divided government dominance. Zero instances of unified control occurred across all seven Congresses—a complete fourteen-year absence unprecedented in American history. Divided control characterized 86% of the era’s Congresses, with opposition Congress accounting for the remaining 14%. This configuration proved remarkably stable. Rather than fluctuating between unified and divided states, the political system remained locked in divided government across three presidential administrations (Carter, Reagan, and Bush). The mechanisms sustaining this pattern—including the divergence between presidential and congressional electoral coalitions—operated with such consistency that unified control appeared structurally blocked rather than merely infrequent. This era demonstrated that divided government could persist not just as an occasional outcome but as an entrenched equilibrium.

Era 6: Polarized Governance (1993-2027) presents the most structurally fragmented landscape in the historical record. While unified control has returned, appearing in fourteen of forty Congresses (35%), its character has fundamentally transformed. The average duration of unified runs has collapsed to two and one third years—less than one-third the Era 1 average and barely half the Era 4 average—with a median of just two years and a maximum run of only four years. Most significantly, opposition Congress configurations have surged to twelve of forty Congresses (30%)—more than quadruple the Era 1 rate and more than double the Era 4 rate. When combined with divided control at 35%, non-unified configurations now characterize 65% of governance. The balance has shifted decisively toward unified Republican control (eleven Congresses) over unified Democratic control (three Congresses), though both parties have experienced the structural challenge of sustaining control beyond a single two-year cycle. This era reveals that the return of unified government after Era 5’s complete absence does not represent a return to earlier patterns, but rather the emergence of a new structural regime characterized by high-frequency oscillation between configurations and the normalization of complete congressional opposition to the president.

This temporal pattern reveals three structural features. First, unified control has not disappeared but has become increasingly episodic and short-lived. Second, the variety of divided configurations has expanded, with opposition Congress emerging as a regular rather than exceptional arrangement. Third, the transition has been neither linear nor uniform—different eras display distinct structural signatures, suggesting that divided government emergence reflects the interaction of multiple causal mechanisms rather than a single deterministic trend.

Explanations for the Rise of Divided Government

The political science literature offers multiple, often complementary, explanations for the emergence and normalization of divided government as a structural feature of American politics. These explanations operate at different levels of analysis and emphasize distinct causal mechanisms, though scholars increasingly recognize their interactive effects.

The foundational explanation emphasizes the deliberate design choices embedded in the Constitution itself. Staggered electoral calendars—with House elections every two years, presidential elections every four years, and Senate elections on six-year cycles with one-third of seats contested biennially—create multiple opportunities for voters to produce split control even absent intentional ticket-splitting. David Mayhew’s seminal book Divided We Govern demonstrated that divided government, far from representing system failure, reflects the normal operation of separated institutions with distinct constituencies and electoral rhythms. Morris Fiorina’s work extended this analysis, arguing that the midterm election pattern—in which the president’s party typically loses House seats—functions as a quasi-constitutional mechanism for checking executive power, even if voters do not consciously intend this outcome.

The staggered Senate presents a particular structural constraint. Because only one-third of Senate seats are contested in any election cycle, maintaining unified control requires sustained electoral success across multiple cycles. A party can win the presidency and House decisively but inherit a Senate controlled by the opposition if previous cycles favored their opponents. This institutional feature compounds the difficulty of achieving—and especially of sustaining—unified control.

A second cluster of explanations centers on the transformation of party competition itself. Systematic research by McCarty, Poole, and Rosenthal documented the dramatic ideological polarization of congressional parties beginning in the late 1970s, with the two parties moving toward ideologically coherent and distinct positions. This polarization has been accompanied—indeed, arguably driven—by partisan sorting among voters, whereby conservatives increasingly identify as Republican and liberals as Democratic, while the centrist middle contracts.

Paradoxically, this partisan sorting initially increased the potential for divided government during a transitional period. As Gary Jacobson has documented, the erosion of split-ticket voting—the practice of voting for different parties for different offices—has made federal elections more nationalized and party-centered. When split-ticket voting was common (declining from over 25% in the 1970s to under 10% by the 2010s), electoral volatility could produce divided outcomes even within single election cycles. The current low-split-ticket environment means that unified government is more likely to emerge from a strong national party wave, but also that such unified control faces immediate vulnerability at the next midterm election.

Related to polarization, the nationalization of congressional elections has transformed midterm election patterns. In Insecure Majorities, Frances Lee argues that parties now operate in a context of perpetual competitive threat, where control of Congress regularly hangs in the balance. This nationalization means that midterm elections increasingly function as referenda on presidential performance, producing systematic swings against the president’s party. When combined with the baseline difficulty of maintaining Senate control across staggered cycles, nationalization creates a strong structural headwind against sustained unified control.

Jacobson’s analysis of the “presidential penalty”—the systematic losses suffered by the president’s party in midterm elections—identifies both mechanical and behavioral mechanisms. Mechanically, the president’s party often defends more marginal seats after winning them in the presidential election wave. Behaviorally, midterm electorates differ from presidential electorates in composition and motivation, with opposition party voters showing higher turnout rates and greater enthusiasm. Subsequent work by Sides, Teixeira, and Vavreck has confirmed that these midterm dynamics persist even in highly polarized environments.

A fourth explanation emphasizes the geographic distribution of partisan support. Jonathan Rodden’s Why Cities Lose documents how Democratic voters’ concentration in urban areas creates representational inefficiencies in both House districting and Senate representation. Because Senate representation is not population-proportional, states with small populations have equal representation to large states. The correlation between state population size and Democratic voting strength means that Democrats face a structural disadvantage in achieving Senate majorities even when they win the popular vote nationally.

Similarly, Bill Bishop’s The Big Sort describes the self-reinforcing process of geographic sorting, whereby Americans increasingly live in politically homogeneous communities. This sorting has made House districts less competitive and more ideologically consistent, reducing the likelihood of ticket-splitting at the district level while potentially creating partisan asymmetries in the translation of votes to seats. Research on redistricting by Carson and Engstrom suggests that while gerrymandering contributes to these patterns, the fundamental geographic concentration of partisans operates independently of redistricting choices.

A fifth set of explanations focuses on transformations in the political information environment. Markus Prior and others have documented how media fragmentation—the shift from limited broadcast networks to abundant cable and digital options—enables partisan selective exposure. Voters can now consume news that reinforces rather than challenges their partisan identities, potentially intensifying polarization and reducing cross-party voting.

Simultaneously, the rise of affective polarization—the phenomenon whereby partisans increasingly dislike and distrust the opposing party—has been documented by Iyengar, Sood, and Lelkes. This emotional dimension of partisanship differs from ideological polarization; voters may not hold more extreme policy positions, but they view the opposing party with greater animosity. Related research on negative partisanship by Abramowitz and Webster shows that voters are increasingly motivated by opposition to the other party rather than by support for their own. This affective dimension may sustain divided government by making voters more willing to check unified control, viewing single-party dominance as threatening even when they support that party in other contexts.

The literature contains ongoing debate about the relative weight and causal sequencing of these factors. Fiorina has argued that elite polarization precedes and exceeds mass polarization, suggesting that institutional and strategic factors may be more fundamental than voter preferences. Others, including Abramowitz, emphasize the reality of mass partisan sorting and genuine ideological polarization among ordinary voters. There is broad consensus, however, that multiple reinforcing mechanisms are at work, and that divided government reflects the interaction of constitutional design, party competition, electoral behavior, and information environments rather than any single dominant cause.

Importantly, scholars distinguish between divided government as an outcome (the mere fact of split control) and divided government as a persistent equilibrium (a stable configuration difficult to exit). The mechanisms that increase the frequency of divided government do not necessarily explain why it endures once established.

Persistence, Reversal, and Structural Constraints

The political science literature distinguishes sharply between explaining the increased frequency of divided government and explaining its persistence once established. While the mechanisms discussed above account for why divided government occurs more often, the question of why it endures—or under what conditions it might be reversed—requires attention to feedback effects, institutional constraints, and the conditions necessary for partisan realignment.

The most immediate structural constraint on unified control persistence stems from the two-year House election cycle combined with the regularity of midterm losses. Sarah Binder’s work on legislative productivity under divided government notes that newly elected unified governments face an immediate clock. They typically have less than two years before facing a midterm election that historically produces losses for the president’s party. This creates a narrow window for legislative accomplishment and limits the time available for unified policy regimes to demonstrate success before facing electoral judgment.

The regularity of this pattern—the president’s party has lost House seats in thirty-seven of forty midterm elections since the Civil War—means that unified control established in a presidential election year faces a high probability of termination at the midterm. Even controlling for economic conditions, presidential approval, and other factors, a systematic midterm penalty persists.

The Senate’s six-year terms and staggered election schedule create a distinct persistence dynamic. A party gaining Senate control inherits electoral maps shaped by outcomes from the previous two cycles—six and twelve years prior. If those earlier cycles favored the opposing party, the new Senate majority must not only win the current election but also defend seats won in those favorable earlier cycles, even if the political environment has shifted.

Frances Lee and Sean Theriault’s work on Senate dynamics emphasizes how this arithmetic creates path dependence. A party that wins unified control after a period in opposition may find that the Senate seats up for election in the next cycle are disproportionately their own seats in vulnerable states, making it difficult to sustain the majority even if their national political position remains strong.

Several scholars argue that polarization itself creates dynamics that sustain divided government. When parties are highly polarized, each party’s coalition becomes more internally homogeneous but also more distinctive from the opposing party. This can increase the perceived risk of unified control—even among voters who prefer one party—because the programmatic distance between parties makes single-party control appear more consequential.

Additionally, polarization reduces the supply of moderate candidates who might appeal across party lines in swing districts or states. As Theriault’s The Gingrich Senators documents, the entry of more ideologically extreme cohorts into Congress can shift the overall ideological distribution of the parties, making bipartisan coalition-building more difficult and increasing the likelihood that voters perceive stark choices rather than nuanced differences.

Keith Krehbiel’s work on pivotal politics suggests that legislative gridlock under divided government creates informational problems for voters. When policy change is difficult, voters may struggle to assign credit or blame clearly, potentially making accountability mechanisms less effective. This ambiguity can advantage opposition parties in midterm elections, as voters dissatisfied with status quo policies may blame the president’s party regardless of whether congressional opposition has blocked that party’s agenda.

Simultaneously, the prevalence of negative campaigning in polarized environments may make voters more risk-averse about unified control. If voters are exposed primarily to negative information about the opposing party, they may prefer divided government as a form of insurance, even if they nominally prefer one party’s policy positions.

The literature identifies specific conditions under which entrenched divided government might give way to sustained unified control. The most frequently cited mechanism is partisan realignment—a fundamental reshuffling of party coalitions that produces a new dominant majority. Burnham’s concept of critical elections and subsequent realignment scholarship suggests that major economic crises or social transformations can trigger partisan sorting that produces extended periods of single-party dominance, as occurred during Era 3 (1933-1947).

However, contemporary scholars question whether classic realignment theory applies in the current era. In Electoral Realignments, David Mayhew argues that the traditional realignment model overstates the frequency and coherence of such episodes, while in Party Position Change. David Karol emphasizes that party coalitions evolve gradually through elite repositioning rather than through sudden voter shifts.

More recent work suggests that contemporary polarization may have altered the potential for realignment. When parties are already highly sorted ideologically, the opportunities for coalition reshuffling may be limited. Abramowitz and Teixeira’s The Emerging Democratic Majority argued that demographic change might produce a new sustained Democratic majority, while others have suggested that Republican advantages in the Electoral College and Senate might produce Republican persistence despite Democratic popular vote majorities. Both arguments, however, have proven premature, suggesting that the structural factors sustaining divided government operate with considerable robustness.

The contemporary literature emphasizes that divided government persistence is conditional rather than inevitable. Circumstances that might facilitate unified control’s return and persistence include (a) a major crisis that produces a clear electoral mandate, (b) demographic or economic changes that shift the political geography of party competition, (c) institutional reforms that reduce representational asymmetries, or (d) elite strategic choices that moderate party positioning and broaden electoral coalitions.

Critically, the literature does not support claims of inevitability in either direction. Just as Era 3’s fourteen-year unified run gave way to Era 4’s mixed pattern, and Era 5’s complete absence of unified control gave way to Era 6’s more varied configurations, the current structural features could themselves prove transient. What the literature does suggest is that absent major disruptions or institutional changes, the mechanisms currently sustaining high frequencies of divided government—polarization, nationalization, midterm dynamics, and representational asymmetries—tend to reinforce one another through feedback effects.

Taken on its own terms, this literature explains why divided government has become more frequent, more varied in form, and more difficult to exit. It also makes clear that fragmentation is neither accidental nor self-correcting, but sustained by reinforcing institutional, electoral, and informational dynamics. What it does not address is how this condition alters the behavior of complex institutional systems operating within it. To move from explanation to implication, the analysis now returns to the seven-edge framework, using it to examine how persistent divided government reshapes the transmission of pressure across domains rather than determining outcomes within any single one.

Divided Government and the Seven-Edge System

Viewed through the seven-edge system, divided government functions less as a legislative impediment than as an operating condition that reshapes how pressure propagates across institutional boundaries. When unified authority is episodic and short-lived, the system does not stall. It reorganizes. Transmission becomes uneven, sensitivity concentrates at specific edges, and responsibility diffuses across domains that are poorly synchronized in time.

The most immediate effect of persistent fragmentation is heightened activity at the system’s boundaries—particularly at Edge 1 (Narrative → Law) and at the political feedback loop spanning Edge 6 (Macro Conditions → Politics) and Edge 7 (Politics → Narrative). The above analysis makes clear why this occurs. Midterm penalties, staggered electoral cycles, and Senate arithmetic sharply compress the window in which unified control can operate. As a result, narratives proliferate even as durable codification becomes harder to achieve.

Under divided government, Edge 1 becomes a site of acceleration but limited realization. Narrative production intensifies, frames polarize, and legal legibility becomes a scarce resource. Yet the threshold for successful translation into law rises. Legislative veto points prevent most narratives from stabilizing into comprehensive statutory regimes, producing what might be described as a surplus of legibility—a growing inventory of investigatory authorities, model statutes, and definitional constraints that circulate without consolidating into durable federal law. The consequence is not narrative dominance, but congestion—high salience combined with low conversion.

At the opposite boundary of the system, divided government amplifies the feedback between macroeconomic stress and political behavior. Edge 6 becomes especially consequential as macro conditions substitute for legislative resolution as the primary driver of political salience. Inflationary pressure, insurance strain, or regional economic dislocation reshape priorities even when their causes are diffuse or contested. These conditions do not transmit precise causal accounts into politics; they compress time horizons and intensify blame competition. Under fragmentation, no single actor can plausibly claim control over outcomes, yet all actors face electoral consequences for aggregate conditions they cannot fully manage.

This pressure is then recycled into narrative production through Edge 7, which becomes increasingly central under divided government. When legislative closure is rare, politics functions less as a means to law than as a demand engine for narrative frames. Electoral signals, partisan media incentives, and primary dynamics authorize and amplify explanations that mobilize attention rather than resolve institutional conflict. The system evolves through selective amplification rather than corrective convergence, reinforcing polarization and sustaining fragmentation.

While the boundary edges become more sensitive, the interior of the system—Edges 2 through 5—exhibits a different pattern. These edges remain active, but they operate primarily as zones of attenuation, lag, and quiet accumulation rather than decisive action.

Legal signals continue to influence capital allocation through Edge 2 (Law → Capital Allocation), but under divided government those signals are often ambiguous, contested, or unevenly enforced. Anticipatory compliance, risk hedging, and selective market participation become dominant responses. Enforcement uncertainty increases asymmetric error costs. The downside of misjudging future liability often outweighs the cost of preemptive, and sometimes contradictory, adjustments. What travels across this edge is constraint and permission filtered through inference, not instruction.

These signals propagate into firm governance through Edge 3 (Capital Allocation → Firm Governance), where boards and managers interpret fragmented cues conservatively. Governance adjustments tend to be procedural rather than transformational—reframing disclosures, reallocating attention, or reducing public salience—while preserving operational flexibility. Under divided government, governance becomes a site of risk management rather than strategic resolution.

The translation from governance intent to real economic activity through Edge 4 (Firm Governance → Real Economy) is further slowed by execution constraints and competitive pressures. Operational decisions reflect recalibrated thresholds and delayed investments rather than abrupt reversals. Effects accumulate gradually, often below the level of immediate visibility.

These incremental adjustments aggregate into macroeconomic conditions through Edge 5 (Real Economy → Macro Conditions), but the aggregation process itself obscures causality. Nonlinearity and measurement lag ensure that macro-outcomes are rarely attributable to specific governance or legal interventions within the time horizon of political accountability. This opacity feeds back into the political dynamics described above, reinforcing blame competition and sustaining divided control.

A further consequence of fragmentation is temporal compression. As the duration of unified control collapses, rare windows of alignment produce high-velocity interventions as governing coalitions attempt to compress extended agendas into brief periods. These interventions can be nonlinear, triggering governance and capital responses that are difficult to reverse once the system returns to divided control. The lag between implementation and macroeconomic effect then distorts accountability. Political judgment occurs before outcomes are legible, while outcomes manifest after authority has shifted.

Taken together, these dynamics produce a system characterized by high resonance at its symbolic boundaries and high impedance at its institutional core. Pressure moves rapidly through narrative and political signaling but encounters friction at the edges that require sustained coordination and temporal continuity. The system remains active, but its capacity for closure is constrained. Influence circulates, but it arrives transformed, delayed, or redirected.

Understanding divided government in this way clarifies why ESG conflict appears both omnipresent and unresolved. Disputes are fought intensely over framing, legitimacy, and blame, while interior adjustments proceed quietly and unevenly. Outcomes emerge not from decisive intervention at a single edge, but from the cumulative interaction of delayed transmissions and amplified feedback at the system’s boundaries. The seven-edge framework makes visible how fragmentation reshapes not only political outcomes, but the architecture of transmission itself.

What This Means for Climate Change Policy

The analysis in this chapter does not point to a single climate outcome or a preferred policy design. It identifies instead a structural condition that now defines the operating environment in which US climate policy unfolds. Divided government, once treated as a temporary interruption to unified authority, has become a persistent feature of contemporary governance. Its significance for climate policy lies not in what it blocks outright, but in how it reshapes the pathways through which climate-related pressure moves, accumulates, and is transformed.

Seen through the seven-edge system, divided government alters the balance between visibility and control in climate governance. Pressure does not dissipate when authority fragments; it is displaced. Climate narratives become more salient, political signaling intensifies, and policy commitments are increasingly framed symbolically, while institutional interiors absorb strain without producing durable settlement. Law becomes harder to stabilize, capital and firm governance adapt defensively, and real-economy climate outcomes emerge slowly and unevenly, often beyond the horizon of political accountability. The system remains active, but its capacity for closure is constrained.

This helps explain a recurring pattern across the four scenario chapters. Climate policy appears simultaneously urgent and unresolved. Ambitious commitments are announced, contested, revised, and re-announced, while underlying emissions trajectories, infrastructure investment, and adaptation outcomes change incrementally or with substantial delay. This is not primarily a failure of political will, administrative competence, or market response taken in isolation. It is a property of a system operating under sustained fragmentation, where climate-relevant decisions are distributed across domains with different logics, time horizons, and constraints.

The seven-edge framework clarifies why climate interventions that appear decisive within one domain often produce muted or delayed effects elsewhere. Legislative action does not translate cleanly into capital reallocation; capital signals do not map directly onto firm-level operational change; and operational adjustments aggregate into macro conditions only over extended periods. Accountability is therefore persistently contested. Climate outcomes emerge from interactions across edges rather than from discrete policy choices within any single arena. When effects become visible, they are frequently attributed to proximate political events rather than to the accumulated transmissions that produced them. This misattribution is not accidental. It is reinforced by the structure of the system itself.

Understanding climate policy in system-level terms does not resolve disagreement about goals, trade-offs, or acceptable risk. It reframes the nature of the challenge. The central question shifts from whether particular climate policies are right or wrong to how climate-related pressure is likely to propagate, stall, or reappear under prevailing institutional conditions. That reframing makes visible both the limits of individual policy interventions and the sources of their unintended consequences.

This perspective also clarifies why cycles of escalation recur in climate politics. When legal closure is rare and macroeconomic feedback is delayed, narrative and political edges absorb and re-amplify unresolved tension. Climate conflict becomes self-reinforcing not because solutions are unavailable, but because the pathways that would produce durable alignment are misaligned with the system’s temporal and institutional constraints. What appears as ideological polarization often reflects structural persistence.

What this analysis ultimately shows is not why climate policy fails, but why it behaves in the particular ways observed. The recurrence of ambitious announcement alongside institutional inertia, the prominence of symbolic confrontation alongside incremental change, and the difficulty of attributing outcomes to specific decisions are not anomalies. They are predictable consequences of a system operating under sustained fragmentation, in which transmission is uneven, timing is misaligned, and no single domain exercises decisive control. The seven-edge framework renders that structure legible, clarifying both the reach and the limits of climate policy action within contemporary US governance.

The shock has occurred with the same sustained visibility that defined the prior scenarios. Insurance markets contract or exit high-exposure regions. Infrastructure systems experience failures that correlate with projected climate stress patterns. Agricultural yields decline in documented geographic clusters. Supply chains adapt to recurrent disruption that no longer registers as exceptional. The economic manifestation is neither abstract nor singular but persistent—a transition from contingent risk to operating condition that appears in earnings guidance, municipal budget constraints, and credit repricing across exposed asset classes.

Unified Democratic control of the Presidency, Senate, and House creates the political configuration most commonly assumed to enable comprehensive climate policy response. The party’s stated positions support climate-risk integration into financial regulation, mandatory disclosure frameworks, and coordinated adaptation investment. Electoral interpretation emphasizes constituent demand for federal leadership on climate impacts. The institutional machinery—legislative majority, executive authority, regulatory appointments, and appropriations control—appears aligned for coordinated action.

But unified control does not eliminate institutional constraint. It concentrates political authority within a single party while simultaneously exposing that party to comprehensive responsibility for every implementation failure, every unintended economic consequence, and every adverse legal precedent. The anti-ESG coalition does not dissolve under Democratic governance. It redirects strategy from legislative blocking to litigation preparation, from appropriations constraint to judicial challenge, and from narrative production emphasizing regulatory overreach to narrative production emphasizing economic harm and constitutional violation. The shock creates pressure for response. Unified control determines where that pressure can be directed and which constraints remain binding despite political alignment.

The system’s constants remain operative. Legal uncertainty persists because statutory authority for climate-specific financial regulation remains contested. Fiduciary duty interpretation divides across jurisdictions and investor constituencies. Liability exposure from aggressive action—regulatory takings claims, preemption challenges, energy sector litigation—creates defensive positioning even within supportive administrations. Firms adapt to regulatory pressure but do so strategically, managing compliance costs and political exposure rather than converging on shared climate-risk frameworks. Capital responds to regulatory signals but discounts durability when judicial reversal probability remains elevated. Most critically, the federal system’s structure ensures that implementation occurs through state and local execution, where political opposition concentrates and where regulatory reach confronts practical limits.

The scenario therefore examines how unified Democratic control processes a visible climate shock through a system where political alignment enables regulatory action but does not eliminate the constraints that govern legal durability, capital allocation, firm behavior, and implementation execution. What changes is the concentration of political accountability. What persists are the institutional frictions that prevent visibility from translating directly into settled policy outcomes. The result is motion at the federal level that encounters friction at the implementation level, producing activity without comprehensive settlement.

The chapter proceeds through the transmission pathways that structure institutional response under unified Democratic control. In system terms, three edges govern how visible shock is processed. First, the Narrative → Law edge (E1) shapes how physical damage and economic disruption are rendered legally legible, translating attribution and urgency into statutory authority and regulatory justification rather than into unconstrained mandate. Second, the Law → Capital Allocation edge (E2) determines how newly asserted authority alters permission structures for investment and supervision, reshaping incentives while transmitting uncertainty alongside direction. Third, the Macro Conditions → Politics edge (E6) concentrates salience within the governing coalition, converting aggregated economic stress into political imperative without resolving disagreement about sequencing, scope, or durability. Together, these pathways explain why unified control accelerates institutional movement while shifting the locus of constraint from legislative blocking to implementation friction and from appropriations conflict to judicial confrontation, producing activity without comprehensive settlement.

Entry Conditions and Political Interpretation

The Democratic Party enters unified control in January 2029 interpreting the preceding election as validation of climate-risk governance and federal coordination on adaptation infrastructure. Campaign messaging emphasized the economic costs of inaction, the failure of market-only approaches to address systemic physical risk, and the need for regulatory clarity to enable capital allocation toward resilience. House and Senate majorities include members from constituencies experiencing direct shock impacts—coastal districts facing insurance market collapse, agricultural regions managing sustained drought, and municipalities confronting infrastructure failure—alongside members from constituencies where climate advocacy has long preceded visible physical harm.

This diversity creates intraparty tension that unified control cannot externalize. A senator from a coastal state experiencing property insurance crisis faces different constituent pressure than a senator from an inland manufacturing state where climate impacts remain less visible. A representative from a district with significant fossil fuel employment navigates different political constraints than a representative from a district demanding immediate emissions regulation. These tensions existed during the campaign but intensify when abstract climate positions must translate into specific regulatory actions with concentrated economic impacts.

The shock’s visibility creates political necessity to act but does not specify which actions are viable within coalition constraints. Progressive members demand comprehensive emissions regulation, mandatory climate disclosure with standardized metrics, and federal preemption of state anti-ESG laws. Moderate members support adaptation infrastructure investment and enhanced financial risk supervision but oppose measures that could be characterized as energy sector punishment or could generate adverse economic impacts in swing districts. Leadership must find legislative and regulatory approaches that satisfy both constituencies without fracturing the governing coalition.

This necessity drives demand for attribution frameworks that connect visible harm to regulatory solutions within the coalition’s ideological boundaries. The shock must be explained in ways that justify federal intervention, assign responsibility for past inaction, and position regulatory action as economic necessity rather than ideological preference. Unlike Republican attribution strategies that deflect from emissions causation, Democratic attribution strategies emphasize the connection between warming and economic harm while carefully managing how responsibility for mitigation and adaptation is allocated across federal, state, and private actors.

Coalition management therefore shapes which regulatory responses advance and in what sequence. Early action concentrates on areas with broad intraparty support—enhanced climate-risk disclosure requirements tied to investor protection, supervisory guidance for financial institutions on physical risk assessment, and adaptation infrastructure appropriations framed as economic development. More contentious measures—carbon pricing, mandatory emissions reduction targets, federal energy sector regulation—face sequential positioning that builds political support through demonstrated success of earlier initiatives while avoiding early confrontations that could fragment the coalition or generate adverse legal precedents.

Agenda effects follow. The shock becomes a forcing event that justifies acceleration of regulatory initiatives previously constrained by divided government. But acceleration occurs within boundaries defined by litigation risk, implementation capacity, and coalition unity. Leadership must balance constituent demand for visible federal action against institutional constraints that make comprehensive climate policy legally vulnerable and politically risky. The resulting pattern is selective ambition—aggressive action in domains where statutory authority is clearest and coalition support is strongest, defensive caution in domains where legal vulnerability is highest and intraparty disagreement is sharpest.

The shock generates pressure for coordinated response. Unified control concentrates authority to direct that response. But the direction chosen reflects not only policy preference but strategic calculation about which actions can survive the constraints that operate regardless of electoral configuration. At this stage, pressure moves primarily through the macro-to-politics edge, as Macro Conditions → Politics (E6) converts aggregated economic stress into political imperative without resolving disagreement about implementation strategy or eliminating constraints on regulatory durability.

Legislative Dynamics under Unified Democratic Control

Congressional Democrats face the opportunity and burden of legislative capacity. Unlike divided government scenarios where messaging legislation substitutes for enacted law, unified control requires producing statutes that can survive presidential signature, regulatory implementation, judicial review, and potential future repeal efforts. This transforms legislative strategy from position-taking to institutional design.

The House advances climate disclosure legislation requiring publicly traded companies to report greenhouse gas emissions, physical risk exposure, and transition planning using standardized methodologies. The legislation ties disclosure requirements to existing securities law frameworks, emphasizing investor protection and market efficiency rather than environmental objectives. This framing serves dual purposes—building intraparty consensus by connecting climate to financial materiality and constructing legal defensibility by grounding requirements in established regulatory authority. Floor debate emphasizes economic arguments about capital allocation efficiency and risk management rather than moral arguments about climate justice or environmental stewardship.

The Senate faces more acute legislative challenges. The chamber’s narrower Democratic majority—often dependent on moderate members from politically competitive states—creates pressure to moderate House proposals while maintaining sufficient ambition to satisfy progressive constituencies and demonstrate response proportional to the shock’s visibility. Senate committees hold hearings that establish legislative intent supporting climate-risk integration while building records that respond to anticipated constitutional and administrative law challenges. Witnesses include financial regulators, institutional investors, actuaries, and climate scientists who can testify to the financial materiality of climate risk and the statutory basis for federal disclosure requirements.

Committee amendments narrow House proposals in ways that increase legal defensibility while reducing political exposure. Disclosure requirements emphasize material financial risk rather than comprehensive emissions accounting. Safe harbors protect good-faith disclosure efforts from litigation. Phase-in periods allow firms time to develop reporting systems. Exemptions for smaller companies reduce compliance burden. These modifications disappoint progressive members but prove necessary for passage while improving the legislation’s prospects in judicial review.

Appropriations legislation funds the regulatory infrastructure necessary for implementation. Agencies receive budget increases for examiner training on climate-risk assessment, technology systems for processing standardized disclosure, and coordination mechanisms between financial regulators and climate science agencies. Appropriations also fund adaptation infrastructure investments—grid resilience, water management systems, and coastal protection—framed as economic development and job creation rather than purely environmental initiatives. This dual framing maintains coalition unity while building implementation capacity.

Legislative oversight shifts from confrontation to coordination. Committees no longer investigate agency failure to act but instead monitor implementation progress, identify resource constraints, and build records supporting regulatory authority. Oversight hearings feature agency leadership explaining how new disclosure requirements integrate with existing frameworks, how supervisory capacity is expanding to address physical risk, and how coordination with state regulators addresses federal system constraints. This oversight serves to demonstrate competence and build public confidence while creating legislative history that supports agency action in subsequent litigation.

But legislative activity also reveals the limits of unified control. Legislation requiring federal preemption of state anti-ESG laws faces opposition from moderate Democrats concerned about federalism principles and from Republicans who warn that such preemption invites constitutional challenge. Carbon pricing proposals divide the caucus between members supporting economy-wide mechanisms and members concerned about concentrated impacts on energy-producing regions. Mandatory emissions reduction targets face resistance from members representing industrial constituencies. Leadership advances only those measures that can secure majority support without fracturing the coalition, deferring more contentious initiatives to potential future legislative sessions.

The resulting legislative record is significant but bounded. Climate-risk disclosure advances through statutory authorization tied to investor protection. Adaptation infrastructure receives increased appropriations framed as economic development. Financial regulators gain explicit authority and resources for climate-risk supervision. But comprehensive emissions regulation, federal energy sector restructuring, and preemption of state anti-ESG laws remain outside the legislative consensus achievable even under unified control. The shock creates political space for action, but coalition management and judicial vulnerability constrain which actions can survive legislative process and subsequent implementation challenges.

Throughout, legislative behavior reflects awareness that statutes create targets for litigation and that adverse precedents can constrain future Democratic governance more than current Republican opposition. This awareness disciplines ambition, produces defensive statutory construction, and emphasizes incremental expansion of existing authority over novel assertions of regulatory power. The result is legislation that enables regulatory action while building defenses against predictable legal challenges. As this legislation stabilizes, it begins to reshape expectations through Narrative → Law (E1), determining which forms of institutional response gain legal legibility and which remain contested.

Executive and Regulatory Acceleration under Political Support

Regulatory agencies under unified Democratic control experience maximum political support for climate-risk integration combined with maximum legal vulnerability to challenges characterizing such integration as ultra vires regulatory expansion. This combination produces agency behavior characterized by aggressive action structured for legal survival—a pattern of strategic ambition rather than unconstrained policy implementation.

The SEC finalizes climate disclosure rules requiring material risk reporting, governance structure disclosure, and forward-looking risk assessment. The rules use statutory hooks from existing securities law, extensive preambles documenting investor demand and market efficiency benefits, and phase-in schedules that demonstrate reasoned decision-making. Agency leadership anticipates immediate legal challenges and structures rules to survive judicial review under frameworks that require clear statutory authorization and cost-benefit justification. The result is disclosure requirements that represent significant expansion of climate-risk attention while remaining narrower than progressive advocacy organizations demanded.

Financial supervisors—Federal Reserve, OCC, FDIC—issue coordinated guidance requiring large financial institutions to integrate climate-related physical and transition risks into enterprise risk management, stress testing, and capital planning. The guidance emphasizes supervisory expectations rather than categorical requirements, allowing firms flexibility in methodology while establishing regulatory attention to climate exposure. This approach avoids formal rulemaking processes that would trigger more extensive judicial review while creating enforceable supervisory expectations through examination findings and enforcement actions. Institutions understand that material failures to address climate risk will generate supervisory criticism and potential enforcement even absent specific rule violations.

Coordination between financial regulators and climate science agencies accelerates. Joint working groups develop standardized climate scenario frameworks for stress testing. Data-sharing agreements allow supervisors access to physical risk models and emissions datasets. Coordinated guidance documents explain how climate factors integrate with traditional safety-and-soundness supervision. This coordination serves both efficiency and legal defensibility purposes—efficiency by preventing duplicative efforts, legal defensibility by demonstrating that climate-risk supervision serves traditional regulatory objectives rather than pursuing environmental policy outside agency authority.

But regulatory acceleration encounters predictable legal challenges and implementation constraints that unified political control cannot eliminate. Energy sector trade associations file immediate lawsuits challenging SEC disclosure rules as exceeding statutory authority and violating First Amendment commercial speech protections. State attorneys general from Republican-governed states challenge financial supervisory guidance as improper climate regulation disguised as prudential supervision. Individual companies contest enforcement actions on grounds that climate risk assessment methodologies lack scientific consensus and impose unreasonable compliance burdens.

These challenges do not immediately halt regulatory implementation but create uncertainty about durability that affects compliance behavior and capital allocation decisions. Firms invest in disclosure capabilities but structure systems to allow rapid modification if rules are invalidated. Investors increase climate-risk attention but maintain analytical frameworks that separate climate factors from core valuation methodologies in case supervisory expectations reverse. The litigation shadow reduces the effectiveness of regulatory action even when that action has political support and initial implementation success.

Judicial appointments provide long-term strategic advantage but insufficient near-term protection. The President appoints judges and agency leaders committed to climate-risk integration, but judicial consideration of specific regulatory challenges occurs on timelines extending beyond the current administration. Appeals processes ensure that major climate disclosure or supervision rules face years of litigation before achieving final legal settlement. During this period, regulatory durability remains uncertain, affecting the intensity and irreversibility of compliance responses.

Implementation also encounters federalism constraints that unified federal control cannot override. Financial regulation requires coordination with state insurance commissioners, state securities regulators, and state banking supervisors. States with Republican governance resist or delay implementation of federal climate-risk frameworks. State insurance commissioners decline to require climate-risk factors in rate approval processes. State securities regulators issue guidance contradicting federal disclosure expectations. This creates implementation fragmentation where firms and investors face different regulatory environments across jurisdictions, reducing the standardization that federal rules were designed to achieve.

The resulting regulatory environment is one of visible federal action encountering persistent legal and implementation constraints. Agencies advance climate-risk integration using available statutory authority and political support. But the integration achieved is partial, contested, and structured for legal survival rather than policy ambition. The pattern reflects rational behavior under binding constraints—maximizing action within the boundaries defined by litigation risk, coalition management, and federal system limitations. As regulatory action stabilizes, even provisionally, it transmits through Law → Capital Allocation (E2), altering the permission structures under which investment decisions occur while maintaining uncertainty about enforcement durability and legal finality.

Capital and Firm Response under Regulatory Pressure and Reversal Risk

Capital allocators and corporate executives under unified Democratic control face regulatory pressure to integrate climate risk combined with elevated uncertainty about regulatory durability, litigation outcomes, and political reversibility. The resulting behavior pattern is anticipatory adaptation without convergent commitment—a posture that acknowledges regulatory direction while hedging against multiple forms of reversal risk.

Large institutional investors—pension funds, mutual fund complexes, and sovereign wealth funds—increase climate-risk analytical capabilities and expand engagement with portfolio companies on disclosure quality and risk management practices. This response reflects both regulatory pressure and recognition that physical shock impacts create material financial exposures requiring attention. Asset managers enhance stewardship policies to address climate governance, support shareholder resolutions requesting enhanced disclosure, and condition voting decisions on board climate-risk oversight quality. These actions demonstrate responsiveness to regulatory expectations while remaining defensible as traditional fiduciary risk management.

But institutional investors structure climate integration to preserve flexibility for regulatory reversal or adverse litigation outcomes. Climate-risk assessments remain supplementary to core valuation methodologies rather than fully integrated. Investment restrictions based on climate factors avoid categorical exclusions that could trigger anti-ESG litigation or could prove economically costly if regulatory support reverses. Engagement emphasizes disclosure improvement and risk assessment rather than emissions reduction targets that might exceed fiduciary authority. The resulting pattern is elevated attention to climate factors within governance frameworks designed to minimize exposure if regulatory environment shifts.

Publicly traded corporations face direct disclosure obligations and supervisory expectations requiring material risk reporting and governance structure documentation. Compliance investments accelerate—expanded sustainability reporting teams, board committee restructuring, enhanced data systems for emissions tracking, and physical risk assessment. This compliance activity is real and resource-intensive, creating organizational change and elevating climate within corporate governance attention.

Corporate response also reflects strategic positioning that hedges reversal risk and manages litigation exposure. Disclosure emphasizes uncertainty and risk factors rather than quantitative commitments that could generate enforcement actions if missed. Forward-looking statements include extensive safe harbor language protecting against securities litigation. Governance structures emphasize oversight and risk management rather than environmental advocacy that could alienate shareholders or invite political retaliation. Firms separate compliance with regulatory requirements from voluntary sustainability commitments, maintaining flexibility to modify the latter if political environment or legal landscape shifts.

Energy sector firms and their financial partners face acute regulatory pressure combined with concentrated litigation capacity. These firms challenge regulatory authority through coordinated legal strategies while implementing minimum viable compliance to avoid enforcement actions. They contest the statutory basis for climate-specific disclosure while providing general risk factor discussion. They challenge supervisory guidance as ultra vires while participating in stress testing exercises. This dual strategy—legal challenge combined with defensive compliance—reflects rational positioning under regulatory hostility. The firms cannot ignore regulatory pressure but can contest its legal foundation while building records for judicial review.

Private capital markets experience spillover effects from public company regulatory requirements without direct supervisory oversight. Private equity firms face investor due diligence expectations for portfolio company climate risk assessment. Private debt markets encounter lender requests for climate risk disclosure and covenant structures addressing physical exposure. This private market pressure reflects both regulatory signaling—investors and lenders anticipate that climate-risk standards will eventually extend to private markets—and physical risk recognition independent of regulatory requirements. But private market responses remain more heterogeneous than public market responses because enforcement mechanisms and standardization pressures are weaker.

Financial institutions navigate regulatory expectations for climate-risk integration while managing coalition resistance from fossil fuel sector clients and Republican-governed state pension systems. Large banks enhance climate-risk stress testing and scenario analysis capabilities in response to supervisory pressure. They expand sustainable finance product offerings and increase lending to renewable energy projects. But they avoid categorical lending restrictions or financing exclusions that would trigger energy sector litigation or state pension divestment. The resulting strategy is selective enhancement—doing more on climate risk in domains where supervisory attention is greatest while avoiding measures that would generate concentrated political or legal opposition.

Insurance companies face the most direct physical risk exposure and the least flexibility in response options. Underwriting losses from climate-related events require premium increases, coverage exclusions, or market exits regardless of regulatory environment. Unified Democratic control enables enhanced coordination with state insurance commissioners on climate-risk pricing and reserve adequacy but cannot eliminate the fundamental tension between pricing risk accurately and maintaining coverage affordability. Some states require insurers to accept climate risk factors in rate setting. Other states prohibit rate increases that regulators deem excessive regardless of actuarial justification. This creates fragmented regulatory environments where insurers face different constraints across jurisdictions.

The capital market pattern that emerges under unified Democratic control is elevated climate-risk attention without settled integration frameworks. Physical risk is too visible to ignore. Regulatory pressure is too intense to dismiss. But reversal risk—from litigation, political transition, or judicial invalidation—remains sufficiently elevated that capital hesitates to treat current regulatory environment as permanent. Investors and firms respond through hedged positioning—increasing climate capabilities and compliance sophistication while maintaining strategic flexibility to adjust if regulatory foundation proves unstable. This hedging is individually rational but collectively suboptimal, as it prevents the convergent risk pricing and coordinated capital reallocation that durable regulatory frameworks might enable.

That transmission operates most clearly through Capital Allocation → Firm Governance (E3). Regulatory pressure and physical risk combine to elevate climate within corporate attention, but governance integration remains provisional. Boards establish oversight committees and modify executive incentives, but these governance changes are structured to be reversible if regulatory expectations shift. What flows across this edge is heightened attention without structural commitment, producing governance activity that acknowledges regulatory pressure while preserving strategic flexibility for future political configurations.

What This Scenario Produces —and What It Does Not

Unified Democratic control following a visible climate shock produces the most aggressive regulatory response examined across the scenario set. Political alignment enables legislative authorization, appropriations support, coordinated agency action, and reduced blocking capacity from opposition party. But aggressive action within a system of binding constraints produces a distinct outcome pattern characterized by provisional integration rather than settled framework.

This scenario produces:

A legally grounded expansion of mandatory climate-risk disclosure.
Federal climate-risk disclosure requirements enacted through securities law frameworks and tied to investor protection rather than environmental objectives, creating legal foundation for standardized reporting while structuring rules to survive judicial review focused on statutory authority and cost-benefit analysis.

An increase in supervisory expectations rather than categorical regulatory mandates.
Coordinated financial supervisory guidance requiring large institutions to integrate physical and transition risks into enterprise risk management, stress testing, and capital planning, implemented through supervisory expectations rather than categorical rules to maintain regulatory flexibility and reduce judicial review vulnerability.

Widespread organizational compliance activity within regulated institutions.
Elevated compliance activity and organizational adaptation within regulated entities, producing expanded sustainability reporting, enhanced board oversight structures, and increased analytical capabilities for climate-risk assessment as firms respond to regulatory pressure and physical risk visibility.

Federal investment in adaptation framed through economic development channels.
Adaptation infrastructure appropriations framed as economic development and job creation, funding grid resilience, water management systems, and coastal protection through legislative processes that maintain coalition unity while building implementation capacity.

Cautious repositioning by capital markets in response to regulatory direction.
Anticipatory capital market repositioning that increases climate-risk attention within investment analysis and corporate governance without converging on standardized valuation methodologies or irreversible strategic commitments, reflecting regulatory pressure combined with elevated reversal risk from litigation and political transition.

Persistent fragmentation in implementation across jurisdictions.
Implementation fragmentation across federal and state jurisdictions as Republican-governed states resist or delay adoption of federal climate-risk frameworks, creating geographic variation in regulatory intensity and reducing the standardization that federal rules were designed to achieve.

Ongoing legal contestation of regulatory authority.
Sustained litigation challenging regulatory authority, disclosure requirements, and supervisory guidance on grounds of statutory overreach, constitutional violation, and inadequate cost-benefit justification, producing extended legal uncertainty that affects compliance intensity and capital allocation durability.

This scenario does not produce:

A settled legal foundation for climate-risk regulation.
Resolution of disagreement about climate-risk regulation’s statutory foundation, fiduciary status, or constitutional boundaries, as unified political control enables regulatory action but cannot eliminate legal contestation or judicial uncertainty about regulatory authority limits.

Uniform integration of climate risk across financial and corporate systems.
Standardized climate-risk integration across financial supervision, corporate governance, and capital allocation, as litigation uncertainty, implementation fragmentation, and reversal risk prevent convergence on shared frameworks despite federal regulatory pressure.

Federal override of state-level resistance.
Preemption of state anti-ESG laws or federal override of state resistance to climate-risk implementation, as coalition constraints and federalism principles limit Democratic willingness to pursue preemption strategies that would invite constitutional challenges and fragment intraparty unity.

Comprehensive federal control of emissions or energy systems.
Comprehensive emissions regulation or federal energy sector restructuring, as coalition management, litigation vulnerability, and economic impact concerns constrain legislative ambition even under unified control, producing climate-risk disclosure and supervision expansion while leaving emissions policy largely outside federal regulatory achievement.

Regulatory durability insulated from political or judicial reversal.
Durable regulatory frameworks immune to political reversal or judicial invalidation, as the structural features that make regulatory action vulnerable—contested statutory authority, judicial skepticism of agency expansion, and federal system constraints—persist regardless of unified political control.

Market-wide agreement on climate-risk valuation standards.
Market consensus on climate-risk valuation methodologies or materiality thresholds, as regulatory frameworks remain provisional and contested, preventing capital from converging on shared approaches to risk assessment and strategic allocation despite elevated attention to physical exposure.

This scenario describes how the system processes a visible climate shock under maximum federal political alignment for climate action. The shock generates pressure and unified control enables regulatory response more aggressive than any alternative political configuration could produce. But the response achieved remains bounded by constraints that operate independent of political alignment—litigation vulnerability shaped by judicial philosophy and statutory ambiguity, implementation resistance enabled by federal system structure, and capital hesitance driven by reversal risk and uncertainty about regulatory durability. The result is significant institutional motion that encounters persistent friction, producing climate-risk integration that is provisional rather than settled and activity that reflects regulatory pressure without generating convergent market response. Unified Democratic control changes what regulatory action is possible but does not eliminate the institutional constraints that prevent visibility from translating directly into durable policy settlement.

It is with the utmost sincerity that I would like to recognize the leadership shown by elected officials in the State of Nebraska in preparing their state for the consequences of climate change. I realize this may come as a surprise to you. While I didn’t know about this award, in a way I’m not surprised. I’ve written in many forums, including my book “An American’s Guide to Climate Change: How America Can Lead and Prosper,” about many great things being done by red states to address climate change. There is quite a distance between the political rhetoric of elected officials coming out of these states and the economics on the ground.

In August 2023 the Nebraska Department of Environment and Energy accepted a $3 million Climate Pollution Reduction Planning Grant from the Environmental Protection Agency (EPA), funded by the Inflation Reduction Act (IRA), and set about writing Nebraska’s first statewide climate action plan. This money was used to produce a 12-point Priority Climate Action Plan (PCAP) which includes establishing a Carbon Intensity Score Registry. This was not a desk exercise. The department spent a year on public engagement—webinars, in-person meetings, consultation with other state agencies, public power districts, agricultural trade groups, environmental organizations, and members of the public. It submitted its PCAP to the EPA on March 1, 2024.

Then it competed for the money to carry the plan out, against nearly 300 applications requesting some $33 billion, and it won one! In July 2024 the EPA selected Nebraska for $307 million, one of 25 awards nationally. The official award came through in October of that year. It is the largest federal grant the department has ever received.

Two weeks after the selection was announced, on August 7, 2024, EPA Deputy Administrator Janet McCabe joined Governor Jim Pillen, state officials, and agriculture representatives at Jacobsen’s Farm in Omaha to mark the occasion. McCabe stated that “EPA is pleased to partner with Governor Pillen and the State of Nebraska to deliver community-driven solutions to address climate change, protect public health, and invest in neighborhoods, family farms, and communities across the Cornhusker state.” In response Pillen commented that “I am grateful to EPA for recognizing the uniqueness of our application and our great state. This grant will strengthen our rural communities, our agriculture industry, and allow Nebraska to fully realize the potential of full value agriculture. This grant signifies what can be done when we all come together for Nebraska and her citizens.” The EPA’s announcement describes the program as advancing the goal of achieving net-zero emissions no later than 2050.

Governor Pillen has a right to be proud. The table below shows the state-level winners. Nebraska got the fifth largest award, after a five-state New England coalition, Illinois, a four-state North Carolina coalition, and Pennsylvania. But it was the highest by far on a per capita basis, at more than three times the next state. Even more impressively, Nebraska’s targeted emissions reductions are the second highest on the list in absolute terms and the highest per capita.

ONE RED

The program has a name. Nebraska calls it ONE RED—Opportunity for Nebraska: Reducing Emissions and Decarbonization. Someone in Lincoln worked at that acronym! I couldn’t help but chuckle a bit about this. It is on the state government website, where it remains. The website describes it as “Nebraska’s first state-wide climate action plans to reduce greenhouse gas (GHG) emissions and other harmful air pollutants in the state” through “high-impact, readily-implemented actions to begin reducing GHG emissions through voluntary actions and financial incentives.” Eight programs are funded under it, including rebates to farmers to replace diesel irrigation pump engines with electric motors, solar canopies over parking lots and cattle feedlots, anaerobic digestion at wastewater treatment plants, energy efficiency for commercial and industrial buildings, biochar, and rural community solar. Not to be laughed at is the excellent 98-page work plan, current as of May 6, 2026, and currently undergoing revision and redesign.

The table below shows the distribution of funds across the eight programs, along with expected emissions reductions by 2030 (the grant expires at the end of 2029) and by 2050. 2050. Keep that date in mind.

The table shows that the largest program of all, at $160,434,689, is the Ag Data Bank and Grants Program, which is comprised of three strategies: a Carbon Intensity (CI) Data Bank, Regenerative Agriculture Incentives, and Precision Agriculture Incentives. The data bank pays farmers to provide their CI scores. The submission of scores makes farmers “eligible to apply for the Regenerative Agriculture Incentives and the Precision Agriculture Incentives programs, which provide grants to initiate regenerative ag practices or purchase precision ag equipment, respectively, which in turn can help producers further reduce their CI score.” Furthermore, “Lower CI scores may mean a farmer can get paid a premium price for their crop through a variety of tax incentives and market opportunities.”

How a farmer’s carbon intensity actually gets measured has not been settled. The Department says it will work with the Governor’s Water Quality and Quantity Task Force—agricultural leaders, state agencies, and industry experts—to “identify and prioritize suitable technology and systems,” and that the same task force “will help determine the parameters of the CI Data Bank Program.” The data bank itself will be run by the Department and a subcontractor selected through a further competitive RFP.

Measuring carbon intensity in agriculture is genuinely hard, and Nebraska is honest that the method is a work in progress. Its methodological starting point is a report the state names as the basis for the program, “Climate Neutrality for Nebraska Agriculture: Benchmarking Current Emissions,” prepared by Blonk Consultants, a Dutch life-cycle assessment firm. So the carbon accounting method underpinning the largest climate program in Nebraska’s history was drafted in the Netherlands. You get the irony.

One housekeeping note on the numbers. The Ag Data Bank’s own program page currently gives rounded figures of $150 million and 28 million metric tons by 2030, while the implementation grant table above gives $160,434,689 and 25,128,828 metric tons. Both are Nebraska’s own pages. I have used the implementation grant figures because they are the original award allocation and because the eight programs sum exactly to the $307 million granted. The program pages appear to reflect the redesign now pending EPA approval.

Getting Expert Help

Dealing with climate change is difficult and many large entities in the private and public sectors look for professional help. Nebraska is no different, and it runs a second federal climate program alongside ONE RED. The IRA authorized the Department of Energy to carry out the Home Energy Rebate Programs, and Sections 50121 and 50122 fund home efficiency and home electrification rebates. Nebraska’s formula allocation across the two is $91,268,349: $45,721,229 for home efficiency and $45,547,120 for home electrification and appliances. Together with ONE RED, that is close to $400 million in federal climate money flowing through a single Nebraska department. The state put the design and implementation of the rebate programs out to competitive bid under RFP 120003 O5.

Deloitte & Touche LLP submitted a proposal dated January 21, 2025, and won. Contract 114994 O4 was signed by Dan Kinsella, Deloitte’s Nebraska Office Managing Partner, on December 13, 2025, and by Jesse Bradley, Director of the Department of Water, Energy and Environment, the following day; the term runs six years from execution, not to exceed September 30, 2031. The award on the face of the contract is $800,900, which buys only Part 1, the program design phase—$440,500 for HOMES and $360,400 for HEAR—while the priced schedule behind it totals $17,975,000 across a four-and-a-half-year scope, including $8.06 million to deliver HOMES rebates and $6.60 million for HEAR. About half the contract’s performance hours go to a subcontractor, DNV Energy Services USA Inc., headquartered in Oslo, which will build and run the rebate technology platform, train and manage the contractor network, determine homeowner eligibility, process the rebates, and manage post-installation inspections. Nice to see how Nebraska is getting help from entities in two European countries.

Language in the proposal and the contract gives a good indication of why Deloitte won. Deloitte is described as the leading consulting firm for energy and environmental consulting services, with over 22,500 energy and utility professionals globally. It offers innovative, client-tested climate action services, and has a track record of aligning state priorities with federal funding.

Two members of the firm’s Climate and Sustainability leadership team are named. Anna Holland, the Program Manager, has over 15 years of experience managing energy transition initiatives, with a background that includes leading state-level HOMES and HEAR design efforts as well as advising the central government of a European country on designing a roadmap for net-zero greenhouse gas emissions across its assets and operations. Jim Staley, serving as Program Design Lead, brings 20 years of energy efficiency and utility experience, having architected HOMES and HEAR program designs and applications across multiple states.

Deloitte was meticulous in describing how it would deliver its professional services. In the pricing assumptions of its bid, the firm states that its services will be performed in accordance with the Statement on Standards for Consulting Services issued by the American Institute of Certified Public Accountants (AICPA) that they will not constitute an engagement to provide audit, compilation, review or attestation services as described in the pronouncements on professional standards issued by the AICPA, the Public Company Accounting Oversight Board (PCAOB) or any other regulatory body; and that Deloitte will therefore express no opinion or any other form of assurance as a result of performing them.

Nebraska was careful as well. Its contract terms require the vendor to certify that no relationship exists which either is, or gives the appearance of, a conflict of interest, and to disclose any actual or perceived conflict together with a proposed mitigation plan for the state to consider. Deloitte certified. Nebraska evaluated the certification and signed the contract.

Deloitte Helps Nebraska with Its Audit Problems

Deloitte’s footprint in Nebraska is vast, as can be seen from the state’s public contracts database. A search by vendor reveals multiple open awards, including a $127 million Department of Health and Human Services data analytics engagement alongside the climate contract discussed above. But the most revealing arrangement is statewide Contract 95281 O4. Procured by price-matching a federal GSA schedule, it began on September 30, 2021, and has since grown across nine amendments and 13 addenda.

To understand why it matters, consider how Nebraska audits itself. Unlike private corporations that hire external accounting firms, Nebraska relies on an independently elected constitutional officer, the Auditor of Public Accounts, to examine the state’s books. (This is the norm across all U.S. states.) In April 2022, that independence produced an embarrassing verdict. The Auditor issued a “Disclaimer of Opinion” on the state’s financial statements for the fiscal year ended June 30, 2021. Declining to vouch for the numbers, the Auditor cited inadequate records, deficient internal controls, unreconciled balances, and more than 110 proposed audit adjustments totaling $4.3 billion.

The transmittal letter at the front of that report acknowledged the scramble. It noted that the State Accounting Division had hired an unnamed external consulting firm mid-stream to rescue the close process, run variance analyses, address audit inquiries, and design internal controls to mitigate reporting risk. The state’s published report never named the firm. Neither did the subsequent year’s report.

But you already know who it is, since I did the big reveal at the start of this section 🐥! Contract 95281 O4 with Deloitte & Touche was executed on September 30, 2021, right as the reporting crisis unfolded, with its first and third Statements of Work earmarked for “NE STATE ACCOUNTING,” totaling roughly $751,000 net. Yet the contract document itself offered zero description of the actual scope.

Deloitte, however, had no qualms about publicizing it. In its subsequent bid for Nebraska’s energy rebate contract, published openly on the state’s purchasing portal, Deloitte showcased its resume: it had stepped in to help State Accounting overhaul the year-end close, resolve lease and IT subscription accounting hurdles, and engineer a roadmap of process enhancements. In fact, Deloitte explicitly credited its own intervention with delivering the state’s first clean, unmodified audit opinion in FY2023 since FY2019.

So while Nebraska never named Deloitte for helping it with its audit, Deloitte named itself in a sales pitch to win more state work—and the information reached the public domain through the contract that pitch won.

Now Here’s the Puzzle

Nebraska is doing great work to lower GHG emissions in the state and has an excellent working relationship with Deloitte. Yet on August 24, 2026, 16 Republican state attorneys general sent a letter to the chief executives of Deloitte, EY, KPMG, and PwC. It runs to 38 pages, carries 269 footnotes, and ends with 38 numbered demands (don’t know if 38 is a coincidence or planned). It was copied to the Chairman of the Securities and Exchange Commission, Paul Atkins, and his Director of Enforcement, David Woodcock. The press release describes it as raising concerns about climate activism. And get this. This letter campaign was led by Mike Hilgers, the Attorney General of Nebraska.

I’ll spare you reading a long and rather rambling letter which comes across to me as an earnest effort to make a number of points by a group that knows very little about what they’re writing about. The foundation for the argument is that the Big Four committed themselves to climate-related disclosure through the Task Force on Climate-related Financial Disclosures, a declaration of support for the International Sustainability Standards Board (ISSB) signed at COP28, and founding membership of the Net Zero Financial Service Providers Alliance. From here the letter goes into any number of flights of fancy about the firms at least creating an appearance that they have adopted an objective external to the audit which compromises their professional independence, that there are potential violations of state professional conduct rules, of the duties of integrity and objectivity, of state unfair and deceptive practices statutes, and of state and federal contract terms.

You get the picture. There’s no proof at all for any of these claims. Reads to me like a kind of “Have you stopped beating your wife?” rhetorical strategy. And for anyone worried about me personally, whenever I visit a doctor, I reassure him or her that I feel safe at home 😍.

It is also worth noting what the letter does not do. It sets no response deadline. A 2023 multi-state letter to Net-Zero Asset ,lliance signatories gave 30 days, which created a date on which something either happened or did not. This letter gives no date, which means it cannot be defied. There is no moment at which non-compliance becomes visible. It simply sits there as an open allegation.

The “analytical” claim of the letter, for which no evidence is ever given, is that climate-related disclosures are typically non-material, that requiring them is speculative and expensive busywork of no use to anyone, and that the Big Four are pushing climate disclosures in order to make more money. The AGs wrap themselves in a populist mantle of protecting farmers and small businesses who will be unfairly burdened by costly climate disclosures. In their words: “The Signatory State Attorneys General won’t sit back while farmers and small businesses are unfairly burdened by the imposition of costly climate-related busy work that the Big Four push and, comparatively, can profit from.” Name me one small farmer who gets an audit from a Big Four firm. Furthermore, thanks to the current SEC, there is no climate disclosure rule in the U.S., and even if there were it would apply to listed companies and wouldn’t touch farmers and small businesses at all.

The letter also either misrepresents or ignores what the standards actually say, which is where it gets genuinely odd. The letter’s benchmark for proper accounting is the Financial Accounting Standards Board (FASB). It approvingly quotes FASB’s Concepts Statement No. 8 on materiality. It quotes FASB on neutrality: information presented without bias, not slanted, weighted, emphasized, de-emphasized, or otherwise manipulated. It quotes FASB on freedom from error. It quotes FASB’s definition of who financial reporting is for: existing and potential investors, lenders and other creditors. Hats off to these red state lawyer AGs for getting the U.S. accounting stuff right.

Then it attacks the ISSB for adopting the same concepts. IFRS S1 requires a complete, neutral and accurate depiction, and its Appendix D says plainly that these qualitative characteristics are taken from the conceptual framework that FASB’s own formulation mirrors. Its definition of primary users—existing and potential investors, lenders and other creditors—is FASB’s words, not a dilution of them. The letter cites the formulation approvingly on one page and treats it as evidence of climate capture a few pages later.

The same thing happens with the auditor’s role. The letter quotes the PCAOB correctly for the proposition that materiality is determined issuer by the issuer. Management decides, the auditor tests whether the evidence supports that judgment, and outsiders don’t get to make the call. Then, from outside every one of these companies, 16 attorneys general determine that a whole category of information is immaterial for public companies as a class. I had a good laugh about this one. Sixteen red state lawyer AGs now see themselves as materiality mavens who know better than the companies they say they are concerned about what these companies should be reporting. Maybe a red state materiality land grab?

And then there is the small matter of which standard they were objecting to. The ISSB has two. IFRS S1 sets out general requirements. IFRS S2 is the climate standard, and it governs everything the letter complains about: emissions, time horizons, projection into the future. The letter attacks the climate standard throughout and never cites a single paragraph of it. Every one of its citations is to the general standard. A reader would not learn from these 38 pages that S2 lets a company explain how its climate disclosures fit into the planning horizons it already uses. Yet the letter complains that the framework demands speculation over an undefined period and, out of context, quotes the ever-quotable anti-ESG SEC Commissioner Hester M. Peirce as saying that “many climate disclosures are high-priced guesses about the present and future.” Except that isn’t quite what they quote. The letter renders it “the present and the future,” and cites it to a footnote pointing at somebody else’s comment letter rather than to Peirce’s own sentence. Go figure. Kinda makes ya wonder about those other 268 footnotes, doesn’t it?

This is almost too delicious for words. Every one of the 16 signatories is a licensed attorney; that is a requirement of the office. Sixteen offices staffed by lawyers produced 38 pages and 269 footnotes attacking a standard, and not one of those footnotes cites it. You don’t need a law degree to read a document you’re complaining about. A high school diploma is enough. Or use AI. I doubt any machine would have made this mistake.

What They Say and What They Do

So what to make of all of this? I’m much more impressed with what Nebraska is doing to adapt to climate change than with what it is saying in its long and silly letter to the Big Four. Although it hardly uses the words climate change at all, Nebraska is making a serious effort to reduce emissions and to enable their agriculture sector to adapt to its reality. Of course, Nebraska is not of one mind. Hilgers didn’t sign the Deloitte contracts and Pillen didn’t sign the letter. But the letter rests on a premise, and the lead signatory’s own state government is busy demonstrating that the premise is false.

The letter says climate-related disclosures are typically non-material. Nebraska publishes emissions projections to 2050, to the individual metric ton, as the public justification for the largest federal grant its environment department has ever received.

The letter says long-horizon climate projection is speculation dressed as analysis and singles out KPMG for setting long-term horizons past 2040 and PwC for a scenario running to 2100. Nebraska’s own table has a column headed “By 2050.”

The letter says measuring and reporting emissions is costly busywork that unfairly burdens farmers. Nebraska is spending $160,434,689 to pay farmers to measure their carbon intensity and report it into a state data bank and considers it the single most valuable thing it can do with the money—52 percent of the grant, and 98 percent of the emissions reductions it expects to achieve.

The letter treats net zero by 2050 as an activist objective so plainly external to any legitimate professional function that association with it is disqualifying. Nebraska competed against nearly 300 applicants for an award under a federal program whose stated purpose is achieving net-zero emissions no later than 2050, won it, and put its governor on a farm with Biden’s deputy EPA administrator to announce it.

The letter demands, at Question 27, that the firms describe their structural separations between audit functions and climate advisory work. Deloitte answered that question in a cost sheet filed with the State of Nebraska eight months before it was asked.

And the letter treats a European hand in climate methodology as evidence of foreign capture. The carbon accounting method underpinning Nebraska’s largest climate program was written by a Dutch consultancy, and the state put it on its website.

A Final Sad Irony

I really am encouraged by what Nebraska is doing to deal with climate change, just as I am about Texas, Florida, and many other red states. At the same time, I have to say that the letters coming out of these red states, while mostly political theater for short-term political benefit, aren’t helpful and have long-term consequences that will make it harder to develop bipartisan policies for climate change. I lay all of this out in my Anti-ESG Ecosystem book, and the core chapter for this discussion is “Letters, Laws, and Lawsuits: How Pressure Enters the System.”

Putting words and deeds together in these red states makes me sad. They are a big part of America, so why are they making it harder on themselves to do what they need to do to deal with climate change? That said, I’m by nature an optimistic guy. Naive as this may sound, over time I think red state citizens will make it known that they don’t appreciate the letters, laws, and lawsuits coming out of their elected officials which are inhibiting America’s ability to adapt to climate change.

The physical shock registers across the economy with the same visibility that drove the previous scenarios—insured losses accumulate, infrastructure fails in documented patterns, and financial statements reflect exposures that actuarial models cannot confidently price. The difference in this configuration is not the shock itself but the political system’s capacity to convert visibility into durable institutional response. Democrats control the Presidency and the House of Representatives. Republicans control the Senate. This division creates a system capable of significant executive action but incapable of legislative settlement.

The Democratic President enters office in January 2029 with a mandate interpretation tied directly to the climate shock’s economic visibility. Campaign messaging emphasized competence in managing physical risk, regulatory clarity for capital markets, and coordination between financial supervision and climate adaptation. The House majority shares this interpretation and possesses appropriations origination authority, investigative capacity, and the ability to pass messaging legislation that builds a legislative record even when Senate passage is blocked. The Senate majority, by contrast, interprets the same election as a mandate to constrain regulatory expansion, protect energy sector interests, and prevent what it characterizes as ideological capture of financial regulation.

This chapter examines how the anti-ESG ecosystem processes this configuration. The core logic is straightforward. Executive agencies face maximum political pressure to act but maximum institutional constraint on the durability of that action. The Senate blocks confirmations, conditions appropriations, builds litigation records through oversight, and signals to regulated entities that enforcement actions face extended legal challenge and potential reversal. The result is not paralysis, but a specific pattern of institutional behavior characterized by defensive incrementalism, strategic fragmentation across agencies, and capital market responses to competing risks rather than settled regulatory expectations.

This configuration therefore produces a specific institutional rhythm. The executive branch accelerates early—because political incentives demand visible response to visible harm—while the Senate works to ensure that early action is narrow, proceduralized, and legally vulnerable. The outcome is a regime of action without settlement, in which agencies behave as if they must show movement while simultaneously preparing for reversal. The anti-ESG coalition does not disappear but rather shifts strategy from offense to institutional defense, using Senate leverage to constrain scope, delay implementation, and build the record for judicial reversal.

Two structural dynamics become especially salient in this configuration, even if they are not named explicitly in ordinary political discourse. First, Law → Capital Allocation (E2) dominates the system’s response. Senate constraint operates primarily by shaping the legal environment in which capital must allocate rather than by dictating outcomes directly. Second, Capital → Firm Governance (E3) becomes conditional and fragmented. Capital sends mixed signals because regulatory durability is uncertain, and those signals translate unevenly into corporate governance, disclosure practices, and strategic investment decisions. In this scenario, visibility is high and institutional activity is intense, yet the system struggles to generate stabilized expectations that would allow capital to converge rather than hedge.

The Incentive Shift: Urgency Meets a Blocking Senate

The Democratic President interprets visibility as political imperative. The shock’s economic documentation creates opportunity cost to inaction. Infrastructure damage, insurance market contraction, and supply chain disruption generate constituency demands for federal coordination. The President’s core agencies—financial regulators, bank supervisors, and pension oversight bodies—receive direction to integrate physical risk into their supervisory frameworks using existing statutory authority. This is not ideological preference but electoral interpretation. The coalition that delivered the Presidency expects institutional response proportional to the shock’s economic visibility.

This executive posture is reinforced by the structural logic of accountability under unified party control of the executive and House. The President can claim leadership through administrative action and can credibly threaten agency heads with political consequences for inaction. The House can amplify that imperative through oversight, messaging legislation, and appropriations baselines. The executive therefore experiences visibility as an instruction to act—not because the system has reached consensus on causation or on mitigation, but because the cost of institutional quiet is politically unacceptable in the face of measurable harm.

The Senate Republican majority interprets the same configuration differently. The election delivered divided government, which Senate leadership characterizes as a public mandate for constraint on executive expansion. This interpretation carries institutional weight. The Senate controls confirmations for all regulatory leadership positions, appropriations amendments for agency budgets, and the oversight process that shapes the record for subsequent litigation. The majority does not oppose all climate-related regulation but insists that regulatory action demonstrate clear statutory authorization, avoid back-door climate policy through financial supervision, and preserve state primacy over energy infrastructure decisions.

This creates a temporal asymmetry in political incentives. The President operates on a four-year horizon with front-loaded action incentives. Early regulatory initiatives signal competence and build implementation momentum before potential Senate shifts in the 2030 midterms. The Senate majority operates on a six-year staggered cycle with back-loaded constraint incentives to delay confirmations, condition appropriations, and build the record for judicial challenge knowing that time dilutes executive momentum and creates opportunity for litigation to mature.

The incentive shift manifests first in personnel. The President nominates regulatory leaders with public track records supporting climate-risk integration. The Senate Banking Committee holds extended confirmation hearings that establish legislative intent opposing what the committee characterizes as regulatory overreach. Nominees either moderate their positions to secure confirmation, face extended delays that leave agencies operating under acting leadership with limited authority, or withdraw in favor of less visible appointments to non-Senate-confirmed positions. This confirmation friction does not block all appointments but signals to agency staff that aggressive regulatory initiatives will face both Senate scrutiny and leadership instability.

The second manifestation appears in appropriations. The House passes funding bills with provisions supporting climate-risk disclosure development, expanded examiner training on physical risk assessment, and coordination funding between financial regulators and climate science agencies. The Senate Appropriations Committee amends these provisions with riders prohibiting specific disclosure mandates, limiting the use of third-party climate analytics in supervision, and requiring cost-benefit analysis for any new reporting requirements. The resulting continuing resolutions or negotiated appropriations bills constrain agency capacity while preserving baseline funding, creating a regulatory environment where action is possible but resource-intensive initiatives face explicit prohibition.

A third manifestation appears in the signaling environment faced by market participants. Firms and investors do not wait for legislation to be enacted to form expectations; they infer durability from the visible balance of institutional power. The President’s push to act signals a high likelihood of near-term enforcement and supervisory pressure. The Senate’s blocking posture signals a high likelihood of reversal, defunding, or judicial defeat. The result is not simply uncertainty in the colloquial sense but a distinct pattern of conditional compliance. Firms respond to immediate enforcement risk while resisting structural changes that presume stable long-term regulatory regimes.

In edge terms, this is the law-to-capital pathway operating at full intensity. The Senate’s constraint posture changes the expected payoff to compliance investment. Capital and firms behave rationally by distinguishing between what must be done to reduce near-term enforcement exposure and what would be done if the regulatory environment were expected to persist across time.

The Senate as Constraint Machine

The Senate’s institutional capacity to constrain executive action operates through four mechanisms that do not require coordination but reinforce through parallel deployment: (a) confirmation leverage, (b) appropriations specificity, (c) oversight record-building, and (d) coordinated litigation signaling.

Confirmation leverage extends beyond regulatory leadership to the full network of positions requiring Senate approval. Financial regulatory agencies operate with multiple board positions, general counsels, and regional leadership roles subject to Senate confirmation. The majority can delay consideration, condition approval on policy commitments, or simply decline to schedule votes, leaving agencies operating with incomplete leadership structures. This does not halt agency functioning but limits the scope of initiatives that require board-level authorization, reduces the political capital available for controversial rulemakings, and signals to career staff that aggressive positions may not enjoy durable leadership support.

Appropriations specificity operates through the detailed instructions that accompany funding bills. Unlike broad policy legislation, appropriations riders can prohibit specific uses of funds with precision that targets particular regulatory initiatives. The Senate can prohibit agencies from finalizing rules on climate disclosure, from hiring additional staff for climate-risk supervision, or from contracting with specific third-party data providers for emissions analytics. These prohibitions carry enforcement mechanisms through agency inspector general oversight and create legal liability for agency staff who violate spending restrictions. The result is a regulatory environment where baseline supervision continues but expansion into climate-risk integration faces resource constraints and legal prohibition on specific methodologies.

Oversight record-building serves a litigation preparation function rather than media spectacle. Committee hearings call agency leadership to explain the statutory basis for climate-risk initiatives, establish legislative history opposing broad interpretations of supervisory authority, and create transcript records that courts can reference in determining congressional intent. Witness questioning focuses on extracting admissions (a) that climate-risk supervision extends beyond traditional safety-and-soundness mandates, (b) that disclosure requirements impose costs without demonstrated investor protection benefits, and (c) that coordination with environmental agencies blurs the boundaries of financial regulation. This record does not block current action but provides litigation plaintiffs with congressional interpretation evidence that challenges agency statutory claims.

Coordinated litigation signaling emerges through formal and informal communication between Senate leadership and the anti-ESG legal network. Committee chairs invite testimony from conservative legal scholars who outline the constitutional and administrative law vulnerabilities of climate-risk regulation. Staff coordinate with state attorneys general on the timing of legal challenges to ensure that cases mature as oversight records develop. This coordination does not constitute direct instruction but creates alignment between Senate constraint strategy and litigation strategy, ensuring that judicial challenges benefit from contemporaneous congressional disapproval evidence.

The cumulative effect of these mechanisms is not obstruction but friction that increases the cost of regulatory action and decreases the expected durability of that action. Agencies can still act but must invest disproportionate resources in legal defensibility, accept narrower scopes to avoid statutory ambiguity, and operate with the knowledge that leadership turnover or appropriations shifts could reverse implemented policies.

This friction has a predictable downstream effect on capital markets. Markets discount regimes that appear reversible. The Senate’s behavior therefore changes not only what agencies can do but what investors are willing to treat as “real” in a durable way. When the probability of reversal is high, compliance becomes tactical rather than strategic. Firms build minimum viable compliance architectures rather than reconfiguring governance and risk systems for long-horizon transformation. Asset managers and lenders treat climate-risk integration as a differentiated service offered to some clients rather than a baseline governance standard imposed across portfolios. Under high-reversal probability, the system produces more documentation but less convergence.

House Dynamics: Messaging, Investigations, and Pressure on Agencies

The House Democratic majority cannot legislate into law under divided government but can legislate into record, investigate agency behavior, and pressure executive agencies to maximize action within statutory constraints. These capacities operate in parallel to create a countervailing institutional force against Senate constraint.

Messaging legislation serves multiple functions beyond immediate passage prospects. The House passes climate disclosure requirements, financial risk integration mandates, and coordination frameworks between supervisory agencies knowing that Senate passage is blocked. This legislation establishes the House’s position for future unified government scenarios, creates campaign material for vulnerable Senate Republicans in states affected by the physical shock, and signals to agency leadership that House oversight will evaluate regulatory action against the standard established in passed-but-blocked legislation. Agency staff interpret this signal correctly. Aggressive action may face Senate constraint, but inaction faces House investigation.

Investigative capacity focuses on documenting regulatory capture by anti-ESG interests and agency failures to integrate documented physical risk. Committee chairs schedule hearings examining (a) energy sector influence on regulatory appointments, (b) the revolving door between conservative legal organizations and agency general counsel positions, and (c) specific cases where supervisors declined to address climate-related exposures despite internal risk assessments. These investigations produce subpoenaed documents, sworn testimony, and public reports that create political cost for regulatory caution. The impact operates through reputation rather than enforcement. Agency leaders prefer Senate appropriations battles to extended House investigations that question their professional competence and institutional independence.

Appropriations origination authority matters even when final bills emerge from Senate negotiation. The House writes initial funding bills with detailed instructions for climate-risk program development, examiner training expansion, and disclosure pilot programs. While the Senate amends these provisions, the House position establishes the negotiating baseline and forces Senate appropriators to explicitly strike specific programs rather than simply declining to fund them. This creates political visibility around Senate constraint that can be costly in states experiencing climate shock-related economic disruption.

Pressure on agencies operates through formal oversight and informal communication. Committee chairs request regular briefings on climate-risk supervision implementation, agency spending on climate analytics capabilities, and the status of disclosure rulemaking. Staff maintain direct contact with career agency personnel to understand where political appointees constrain action and where statutory ambiguity creates implementation flexibility. This pressure does not override Senate constraint but ensures that agencies cannot use Senate opposition as excuse for comprehensive inaction. The House expects maximum deployment of existing authority even when statutory expansion is blocked.

The combined effect creates what agency leadership experiences as competing oversight regimes. Senate oversight demands evidence of statutory constraint, cost-benefit justification, and avoidance of climate policy through financial regulation. House oversight demands evidence of risk integration, supervisory competence, and aggressive use of existing authority. Agency responses fragment across this division. Some initiatives proceed with narrow statutory hooks and extensive legal preambles. Others advance through guidance rather than rules. Still others emerge through selective enforcement that builds precedent without formal rulemaking.

This House pressure has a second-order capital effect. House investigations and messaging bills increase the probability that a future unified Democratic government will treat climate-risk integration as unfinished business. That possibility pulls in the opposite direction from Senate reversal risk. Capital therefore receives competing durability signals. Near-term rules are fragile, but longer-term political trajectories may return to more aggressive regulation. For long-lived investments, this produces hedging rather than alignment. Investors do not adopt one baseline; they diversify across scenarios. The resulting pattern is activity without convergence—more investment in analytics, less investment in irreversible strategy.

Executive and Regulatory Behavior under Cross-Pressure

Agencies operating under competing political pressures develop behavioral patterns characterized by defensive incrementalism, jurisdictional fragmentation, and strategic use of process to manage legal vulnerability. The result is regulatory action that acknowledges physical risk without creating durable integration frameworks.

Defensive incrementalism manifests in the scope and pacing of regulatory initiatives. Agencies decline to pursue comprehensive climate-risk disclosure rules that would face immediate Senate appropriations prohibition and extended litigation. Instead, they advance narrow requirements tied to existing disclosure obligations: enhancing management discussion of material risks, requiring board committee climate-risk oversight documentation, and expanding scenario analysis requirements in stress testing frameworks. Each initiative uses statutory hooks from existing securities law, banking supervision authority, or fiduciary responsibility frameworks. The narrowness limits impact but increases legal defensibility.

Preambles to proposed rules expand dramatically in length and statutory citation density. Agencies anticipate judicial review under frameworks that require clear congressional authorization and reasoned decision-making. Proposed rules therefore include extensive documentation of the statutory basis for climate-risk consideration, economic analysis of disclosure costs and investor protection benefits, and response to anticipated legal challenges. This documentation serves dual purposes. First, it demonstrates agency seriousness to House oversight. Second, it provides the record necessary to survive judicial review. The cost is implementation delay and rule complexity that reduces accessibility for smaller regulated entities.

The choice between guidance and rules becomes strategic rather than merely procedural. Rules create enforceable obligations but trigger formal rulemaking processes that include Senate oversight opportunities, cost-benefit requirements, and judicial review vulnerability. Guidance avoids these triggers but creates uncertainty about enforcement discretion and provides weaker basis for supervisory action. Agencies split the difference. Using guidance for principles and frameworks that signal supervisory expectations while reserving rules for specific disclosure items with clear statutory support. This creates a two-tier system where large, sophisticated firms respond to guidance expectations while smaller entities face uncertainty about enforcement thresholds.

Selective enforcement emerges as the primary tool for building climate-risk precedent without formal rulemaking. Supervisors target enforcement actions at documented cases where firms failed to disclose material climate exposures subsequently realized in financial results. These actions establish that climate risk falls within existing materiality frameworks without requiring new disclosure rules. The selection criteria emphasize legal defensibility—clear causation between undisclosed risk and investor harm, public statements demonstrating firm awareness, and losses large enough to justify enforcement costs. This produces climate-risk precedent in narrow fact patterns rather than systematic integration.

Jurisdictional fragmentation occurs because agencies lack coordination mechanisms that Senate appropriations would fund. Different financial regulators develop non-aligned climate-risk frameworks—bank supervisors focus on credit risk from collateral exposure, securities regulators emphasize disclosure materiality, and insurance supervisors concentrate on underwriting and reserve adequacy. This fragmentation reflects both statutory boundaries and strategic calculation that coordination creates litigation vulnerability by suggesting climate policy rather than safety-and-soundness supervision. Firms consequently face overlapping but inconsistent requirements across regulatory jurisdictions.

The behavioral pattern that emerges is action designed for legal survival rather than comprehensive integration. Agencies demonstrate to House oversight that they take climate risk seriously through guidance, selected enforcement, and supervisory statements. Simultaneously they demonstrate to Senate oversight and anticipated judicial review that specific actions (a) rest on narrow statutory authority, (b) address traditional financial risk categories, and (c) avoid policy determinations beyond agency expertise. The resulting system produces documentation of climate risk without creating settled expectations about how that risk integrates into capital allocation decisions.

This is also a classic mechanism of imperfect transmission in the seven-edge architecture. Agency action is occurring at the legal layer, but it is not arriving in capital markets as a single stable instruction. What crosses from law to capital is not “climate risk must be priced in X way,” but “some climate-risk expectations will be enforced sometimes, for some actors, through some channels, and may be reversed.” That is enough to create compliance investment, but not enough to produce standardized valuation or consistent governance convergence.

Capital and Firm Strategy under “Action That Can’t Settle”

Investors and corporations confront a regulatory environment where multiple competing risks operate simultaneously—physical risk from shock impacts, enforcement risk from selective agency action, litigation risk from both disclosure and non-disclosure, and political retaliation risk from positioning that antagonizes either Senate Republicans or House Democrats. Strategic responses fragment across firm size, sector exposure, and shareholder composition.

Large publicly traded corporations with significant institutional ownership increase disclosure sophistication without committing to specific mitigation targets. They enhance climate-risk discussion in financial filings using language that satisfies securities law materiality standards while avoiding quantitative commitments that could trigger enforcement if missed. They establish board committees with climate-risk oversight but structure charters to emphasize fiduciary duty and risk management rather than environmental objectives. They participate in voluntary disclosure frameworks that demonstrate stakeholder engagement without creating enforceable obligations. This strategy manages competing risks by providing documentation for investors concerned about physical exposure while maintaining flexibility to adjust if political environment or litigation landscape shifts.

Financial institutions face particularly acute cross-pressure because their regulatory exposure spans multiple agencies with non-aligned expectations. Large banks enhance climate-risk stress testing capabilities in response to supervisory pressure but avoid formal climate-risk metrics in capital requirements pending clearer statutory authorization. They develop internal methodologies for assessing collateral climate exposure but maintain these as risk management tools rather than lending restrictions to avoid Senate-triggered investigations into credit discrimination. They expand sustainable finance product offerings to meet investor demand while ensuring that marketing materials emphasize financial return metrics rather than climate impact to manage litigation risk from both greenwashing and anti-ESG claims.

Energy sector firms and their financial partners deploy defensive positioning that emphasizes legal compliance and physical risk management while contesting regulatory expansion. They participate in disclosure processes but challenge the statutory basis for climate-specific requirements through coordinated litigation. They document adaptation investments and efficiency improvements while opposing disclosure rules that treat emissions as distinct from other operational risks. They maintain relationships with both Senate Republicans who provide constraint leverage and House Democrats who control investigative capacity, navigating between protection from regulatory expansion and vulnerability to oversight investigations.

Small and mid-sized firms experience the uncertainty as compliance burden without strategic flexibility. They lack resources for sophisticated risk assessment, face conflicting signals about enforcement priorities, and cannot afford the legal analysis required to distinguish mandatory requirements from supervisory guidance. Many adopt minimal disclosure strategies: acknowledging climate risk in boilerplate language without quantification or forward-looking statements. This approach manages litigation risk from disclosure errors while accepting that minimal disclosure may not satisfy supervisory expectations if enforcement actions expand. The result is disclosure divergence where firm size predicts climate-risk integration depth more than actual exposure levels.

Institutional investors fragment across three strategies that reflect different interpretations of fiduciary duty under regulatory uncertainty. Some integrate climate risk as material financial factor and pressure portfolio companies for enhanced disclosure, viewing physical risk as requiring attention regardless of regulatory settlement. Others maintain climate factors in engagement frameworks but avoid voting or divestment actions that could trigger anti-ESG litigation or political retaliation, treating climate as one risk among many without special status. Still others explicitly reject climate integration as speculative and legally risky, viewing agency pressure as beyond statutory authority and likely to face judicial reversal. This strategic fragmentation reflects genuine disagreement about how fiduciary duty intersects with unsettled regulation rather than ideological positioning.

Private capital markets experience spillover from public company regulatory pressure without direct supervisory oversight. Private equity firms face investor due diligence questions about portfolio company climate exposure without clear frameworks for disclosure or valuation adjustment. Private debt markets struggle to price climate risk premiums without comparable data or regulatory guidance on assessment methodologies. This creates information asymmetry between sophisticated investors with climate analytics capabilities and smaller market participants operating without regulatory floors on disclosure.

In this scenario, the capital market response has a distinctive shape. Physical risk is too visible to ignore. Executive pressure to act is too visible to dismiss. Senate constraint is too powerful to discount. The result is a regime of risk recognition without convergence. Nearly everyone acknowledges climate-related physical exposure as relevant, but fewer actors are willing to embed that exposure into durable valuation and governance structures, because the institutional signals are conflicting and the liability landscape is unsettled.

This is where the “parity” point matters across scenarios. Capital is not a passive mirror of politics; it is an optimizing system that responds to the interaction of legal authority, enforcement probability, litigation risk, and time horizon. In a configuration where executive action is likely, but durability is uncertain, capital does not become more decisive. It becomes more conditional. It invests in optionality—analytics, disclosure language, and modular governance structures—rather than committing to long-horizon reallocation that presumes stable rules. Capital behaves rationally. The system outcome is suboptimal.

The capital market pattern that emerges is risk-aware but strategy-fragmented. Virtually all significant market participants acknowledge physical risk as requiring attention. But acknowledgment does not produce consistent integration because regulatory uncertainty leaves multiple defensible strategies available. Firms and investors choose positions based on shareholder composition, sector exposure, and litigation risk tolerance rather than converging on common approaches. The result is increased disclosure volume without increased disclosure comparability, more climate-risk discussion without settled valuation methodologies, and more strategic positioning without market consensus on what prudent climate-risk management requires.

What This Scenario Produces —and What It Does Not

This configuration—Democratic President and House, Republican Senate—generates specific institutional outputs that distinguish it from both unified government scenarios and alternative divided government configurations.

This scenario produces:

Regulatory action without legislative settlement.
Agencies issue guidance, pursue narrow rulemakings, and expand supervisory attention to physical risk using existing statutory hooks. The activity is real and visible, but it lacks the legislative backbone required for durability.

Defensive incrementalism as the dominant agency posture.
Rather than comprehensive frameworks, agencies prioritize legally survivable steps—narrow disclosure enhancements, extended preambles, and selective enforcement tied to clear materiality failures—because Senate constraint and judicial vulnerability make broad mandates high-risk.

A constraint-driven Senate strategy that reshapes durability expectations.
Confirmation delay, appropriations riders, oversight record-building, and litigation signaling do not stop action entirely, but they increase the cost of action and lower the expected longevity of what is implemented.

Competing oversight regimes that fragment institutional behavior.
House Democrats pressure agencies toward visible risk integration; Senate Republicans pressure them toward statutory minimalism and cost justification. Agencies respond with fragmented tools—rules in narrow zones, guidance elsewhere, and enforcement in selective fact patterns.

Capital-market risk recognition without convergence.
Capital treats physical risk as real and increasingly salient, but discounts regulatory durability. Investors and firms respond through hedging, optionality, and disclosure sophistication rather than through standardized valuation and stable governance convergence.

Disclosure expansion without comparability.
Large firms increase narrative sophistication and governance signaling; smaller firms rely on minimal compliance. The system produces more information but not the harmonized comparability that would support consistent pricing across sectors and firm sizes.

This scenario does not produce:

Mandatory standardized climate disclosure with clear enforcement permanence.
Senate constraint, litigation risk, and appropriations limits prevent the emergence of durable, standardized disclosure regimes.

Integration of climate risk into bank capital requirements or insurance reserve calculations.
Agencies avoid the most structurally consequential moves—capital rule integration, reserve mandates, and system-wide coordination—because those moves trigger maximal Senate resistance and judicial risk.

Coordinated inter-agency frameworks that make climate-risk integration routine.
Coordination appears as the most litigation-vulnerable posture: it looks like policy. Agencies avoid it, producing fragmentation instead.

Durable regulatory infrastructure likely to survive political transitions or judicial review.
Even where action is implemented, it is structured for survival, not for permanence. Markets price that fragility.

Market consensus on climate-risk valuation methodologies or materiality thresholds.
Without stable standards and durable enforcement, valuation remains plural and contested. Capital allocators adopt divergent strategies that are each defensible under uncertainty.

Clear boundaries between legitimate financial risk supervision and prohibited climate policy.
The system remains trapped in boundary conflict. Senate strategy is built to keep that boundary contested; House pressure is built to push it. Neither produces closure.

The scenario’s signature characteristic is regulatory action designed for legal defensibility rather than comprehensive integration. Agencies acknowledge climate risk and take visible steps to address it, satisfying immediate political pressure from Democratic constituencies and House oversight. But the steps taken are narrow in scope, extensively documented in statutory justification, and structured to survive Senate constraint and judicial review rather than to create systematic risk integration. This produces movement without settlement. The regulatory system shifts toward greater climate-risk acknowledgment while leaving fundamental questions about scope, methodology, and durability unresolved.

For the anti-ESG ecosystem, this configuration represents a holding pattern rather than defeat or victory. The coalition cannot prevent all regulatory action but successfully constrains the durability and scope of that action through Senate leverage. The resulting regulatory environment creates litigation opportunities, fragments market responses, and prevents the emergence of standardized frameworks that would make climate-risk integration routine business practice. The system processes visibility by producing action, but the action produced does not resolve the underlying disagreement about whether climate risk belongs within financial regulation’s domain.

Under this configuration, political engagement reallocates pressure across institutions without dissolving the constraints that govern regulatory durability, capital allocation, and firm behavior.

The music was already loud when the doors opened. Not nightclub loud. Precisely calibrated. Loud enough bass to be felt through the carpet, enough rhythm to create forward motion, enough familiarity to suggest optimism without producing any particular memory. The kind of music designed to make important people walk slightly faster while believing they had chosen to do so.

Five thousand invited guests were filing into the Grand Meridian Hall at the Elysium Resort in Las Vegas for the opening session of Alignment 2030, Novyra’s annual gathering. It was not a conference. Novyra had stopped calling it a conference years ago. It was an Alignment. Across the enormous, curved screen behind the stage, phrases appeared and dissolved.

HUMAN JUDGMENT AT MACHINE SPEED

ALIGNMENT IS ACTION

THE FUTURE DOES NOT WAIT

Then:

ALIGNMENT 2030

Novyra had been founded ten years earlier. Its market capitalization had crossed one trillion dollars the previous spring. What exactly Novyra did was more difficult to explain. The company described itself as the world’s leading provider of human-machine coordination infrastructure, supplying autonomous-agent orchestration, enterprise cognition, sovereign AI systems, and what it had recently begun calling “decision architecture” to corporations and governments in 83 countries. No one outside Novyra knew exactly what decision architecture meant. Within six months of Novyra introducing the term, everyone in business seemed to be using it.

The audience included Novyra’s top three levels of management; CEOs of its largest customers; leaders of its principal platform partners; pension and sovereign wealth fund executives; global asset managers; cabinet ministers; regulators; senior military officers; university presidents; and a scattering of people whose titles were sufficiently complicated that nobody was entirely certain whom they represented. These were not people easily impressed by celebrity. Most had spent the previous evening being introduced to people described as legendary, transformative, visionary, iconic, or historic.

At precisely 9:00 a.m., the music stopped. Nika Mirov walked alone onto the stage. Mirov was forty-two years old and had founded Novyra at thirty-two. She wore a charcoal suit over a white shirt, with no visible jewelry except a narrow silver band on one wrist. Her dark hair was short, with the first traces of gray showing at the temples.

She moved very little when she spoke. This was one reason people listened. Mirov was famous inside Novyra for sitting silently through twenty-minute presentations, asking a single question, and rendering the preceding nineteen minutes unnecessary.

Her English was flawless, although there remained the faintest trace of an accent journalists had spent years unsuccessfully trying to place. Profiles invariably mentioned that she had “grown up between systems,” a phrase Mirov had used once in an interview at the age of thirty-three and never explained.

“Good morning.” The room quieted. “Every year, we come together to ask what leadership will require next.” The screen behind her went black. “For ten years, Novyra has been built around a simple belief. Intelligence is not enough. Intelligence must be organized, directed, governed, and ultimately turned into action.”

She paused. “This morning, we begin with someone who has spent his life thinking about exactly that.” Something in her expression changed. Only slightly. But people who knew Nika Mirov noticed. She was excited. “I have had the privilege of meeting presidents, prime ministers, scientists, entrepreneurs, artists, and leaders of some of the most consequential institutions in the world.”

A photograph appeared behind her. Robert Rajgopal. Black background. Dark eyes. Slight gray at the temples. No tie. No smile. Underneath:

ROBERT RAJGOPAL

Then:

RR

“His friends call him RR.”

She turned back to the audience.

“The rest of us call him the greatest management thinker of all time.” No one laughed. Mirov was not known for hyperbole. That was what gave the sentence force. A few people who had been checking messages put their devices down.

“RR has written twenty-seven books, nine of them international bestsellers. His work has appeared in forty-three languages. He has published more than four hundred articles and essays. He has advised more than half of the world’s one hundred largest companies, along with presidents, prime ministers, central bankers, asset owners, asset managers, universities, and international institutions.” The screen moved quickly through book covers, journal covers, photographs, television appearances.

“He is the founder of the Rajgopal Institute for Managerial Praxis, chairman of the Global Council on the Future of Management, and founding editor of The Praxis Review. His weekly television program reaches more than forty million people. His twice-monthly podcasts are downloaded twelve million times each month. His daily essay is read by leaders in nearly every country on earth.”

Daily essay.

“RR is also known for something harder to quantify. He maintains personal relationships with thousands of leaders around the world. CEOs tell me he somehow knows when to call before they know they need the call. Young managers write to him and receive thoughtful replies. He reviews research. He comments on manuscripts. He mentors scholars. He advises boards.”

Mirov allowed herself a small smile. “I do not know how he does it.”

Outside the ballroom, five thousand identical black boxes were stacked in perfect rows. Inside each was a copy of RR’s new book.

“Today is especially meaningful because RR has chosen Alignment 2030 for the global launch of his newest work.” The cover appeared.

“I have read it twice,” Mirov said. “I believe it is the most important book RR has written. And I believe it is the book this moment requires.” She looked out across the room. “Every person here will receive a copy today as my personal gift.”

She paused. “RR has personally signed all five thousand.”

Each signature read simply: RR. Each was perfect.

Mirov continued. “At moments like this, we search for clarity. For someone who can tell us not simply what is changing, but what does not change.” She turned toward the wings. “Robert Rajgopal has spent more than twenty years doing exactly that.” Then she stepped away from center stage. “It is my extraordinary privilege to introduce a scholar, teacher, advisor, author, colleague, and friend whom I believe history will remember as the defining management thinker of all time.”

“Robert Rajgopal.”

The lights went out. Completely. No music. No screen. No light from the stage. The room waited. The darkness lasted long enough to become intentional. Then, after almost a minute, a single white spotlight appeared at center stage. Robert Rajgopal was already standing there. No one had seen him enter.

He was tall and lean, mid-forties, with just enough gray at his temples to suggest experience without age. He wore black jeans and a fitted, short-sleeved black technical shirt which revealed his well-muscled physique. No jacket. No tie. No notes. No lectern. No visible microphone. No clicker. The slides, whenever they appeared, simply appeared.

RR waited until the room was entirely still. When he finally spoke, his voice was softer than anyone expected. Low. Measured. Almost musical.

“Thank you, Nika.” He looked down briefly. “That was far too generous.” He paused. “But thank you.”

He looked across the room. “I want to begin somewhere very different from where Nika ended. None of the ideas I am going to share with you this morning began with me.” He took one slow step forward. “I stand this morning on the shoulders of giants.” Behind him appeared three words.

“For more than a century, remarkable men and women have tried to answer three questions. How do we decide where to go? How do we bring people with us? And how do we use the tools of our age to get there? The vocabulary changes. The questions do not.”

A photograph appeared of an elderly man standing beside a blackboard covered with boxes and arrows. “Lucien Varga.” A few people recognized the name. Most did not. “Varga taught an earlier generation how to think about competition. His great contribution was Competitive Geometry: customers, competitors, capabilities, and context.” Four circles appeared. “He gave executives a language with which to see their world differently.”

RR paused. “For his time, it was a profound achievement.”

The photograph changed. A silver-haired woman stood before an auditorium. “Celeste Arden.” Five book jackets appeared behind her.

The Authentic Leader
The Connected Leader
The Vulnerable Leader
The Adaptive Leader
The Post-Authentic Leader

“Arden devoted nearly fifty years to leadership. It is tempting, looking backward, to see contradiction in the evolution of her thinking.” RR shook his head. “That would be unfair. She understood that leaders must speak the language required by their moment.”

A new photograph. “Theo Kestrel saw earlier than most that innovation would escape the R&D department.” Behind him appeared:

The Innovative Enterprise
The Open Enterprise
The Connected Enterprise
The Adaptive Enterprise
The Intelligent Enterprise
The Autonomous Enterprise

“Each formulation belonged to its time. And each helped managers believe they were preparing for what came next.” The audience was quiet. Some were listening closely. Others had begun to wonder where RR was going. The greatest management thinker alive seemed, for the moment, to be giving them a remarkably elegant history lecture. RR appeared entirely unconcerned.

“And then there was Adrian Sorel.” Six covers appeared.

Leading Change
Leading Disruptive Change
Leading Continuous Change
Leading Accelerating Change
Leading Exponential Change
Leading Beyond Change

“Sorel spent forty-seven years studying organizational change.” RR looked toward the covers. “I met Adrian only twice. Both conversations stayed with me. He understood something essential. Managers rarely change because someone proves to them that the world has changed. They change because someone gives them a way to describe the change.”

RR turned back toward the audience. “These thinkers gave generations of managers words. Categories. Models. Frameworks. Ways of making an uncertain world temporarily intelligible. And we should be grateful to them.”

He paused. “But gratitude is not the same as permanence.” Something in the room shifted. RR continued. On the screen appeared:

“I began my own work by reading people like Varga, Arden, Kestrel, and Sorel. Then I began reading the people they had read.” Old photographs appeared. Factories. Telephone switchboards. Offices filled with desks. Executives in dark suits standing beside machinery. “I went backward. Before the internet. Before the computer. Before television. Before commercial aviation. And I encountered something curious.”

A phrase appeared:
THE OLD ORGANIZATION CAN NO LONGER SURVIVE

“That was written almost a century ago.”

Then, one after another:
HIERARCHY MUST GIVE WAY
TECHNOLOGY IS CHANGING EVERYTHING
THE PACE OF CHANGE IS ACCELERATING
LEADERS MUST EMPOWER THEIR PEOPLE

RR walked slowly toward the front of the stage. “The phrases were different. The diagrams were different. But the diagnosis was remarkably stable. The world is changing faster than ever before. The future is increasingly uncertain. Existing institutions cannot keep up. Hierarchy is too slow. Technology is redistributing power. People must be empowered. Organizations must become more adaptive. Leaders must transform themselves.”

RR stopped. “For a century, each generation of management thinkers believed it had discovered a new world. And each developed the framework appropriate to that discovery.” The room was fully attentive now.

RR looked back at the faded photographs. “I say this with humility because I owe these thinkers an enormous intellectual debt. But perhaps the greatest respect we can show our predecessors is to understand the limits of what they could see. Lucien Varga understood strategy as deeply as anyone could have understood it in 1983. Celeste Arden understood leadership as profoundly as anyone could have understood it in 1997. Theo Kestrel understood innovation as brilliantly as his era permitted. And Adrian Sorel understood change for almost half a century.”

He paused. “They were not wrong. They were early.”

He paused again. “Very early.”

Several heads came up at once. The photographs disappeared. “They sensed conditions that were beginning to emerge. They could name them. But they never fully experienced them. They mistook the first tremors for the earthquake.” The screen went black. “And this is where I must part company with the giants on whose shoulders I stand.” For the first time, RR’s voice carried something other than reflection.

Certainty.

“They believed their moment had arrived. It had not. Ours has.”

The screen filled with four words.

“For the first time in the modern history of management, the conditions management thinkers have been predicting for a century are no longer approaching. They are here.”

More than eight and a half billion people inhabited the planet, RR explained. Some societies were aging rapidly while others struggled to create opportunity for overwhelmingly young populations. Climate change was no longer a forecast but an operating condition. Supply chains once designed for efficiency were being redesigned for security. Energy, semiconductors, critical minerals, food, data, technology itself—all had become questions not merely of economics but of national power.

“Corporate decisions increasingly have consequences once associated only with governments.”

Images moved behind him: legislatures, demonstrations, markets, cargo ships, drought, floodwater.

“Many leading democracies are deeply polarized. Trust in institutions has declined. Trust in experts has declined. Trust in media has declined. Trust in business has declined. And yet we have never depended more on complex institutions and specialized knowledge.”

RR stopped. “All of this would be enough to make our moment different.”

He paused. “But it is not what makes our moment unprecedented.”

A single phrase appeared.

“For centuries, technology amplified human muscle. Then it amplified human communication. Artificial intelligence is different. It amplifies cognition.”

The word appeared:

COGNITION

“For the first time, we are building technologies that do not merely execute human decisions. They participate in making them. They analyze. They diagnose. They design. They persuade. They formulate strategies. They evaluate strategies. They coordinate with one another. And increasingly, they act.”

The screen changed.

AGENCY

“This is the threshold. Intelligence has become scalable. Expertise has become replicable. Agency has become technological.” RR looked across the room. “No one here is asking whether artificial intelligence matters. You have already answered that question. You have invested in it. Deployed it. Built businesses around it. Many of you are winning because of it.”

He paused. “And almost none of you feels caught up.” This landed differently. For the first time, RR seemed to be speaking not about the future, but about everyone in the room. “The systems you deployed eighteen months ago are being overtaken by the systems you are deploying today. Capabilities that created advantage last year are becoming expectations this year. The question is no longer whether AI will transform your organization. The question is whether your organization can keep transforming as fast as AI does.”

By now, RR had the room. He began moving again.

“AI can democratize access to expertise while concentrating unprecedented power in the institutions that control the models, data, compute, and infrastructure. It can create extraordinary new opportunities while concentrating extraordinary new wealth. It can make learning more individualized, adaptive, and available—and make it easier for human beings to avoid the struggle through which learning actually occurs. It can strengthen human judgment. And make the abandonment of judgment irresistible.”

He paused. “It can make organizations smarter. And make responsibility harder to locate.”

RR let the sentence hang. “And there is another problem. AI strategy is not a destination. It is a continuous act of reframing. The organizations that win will not be those that chose the right model in 2028. They will be the organizations capable of questioning what worked in 2029 before it becomes the thing that traps them in 2030.”

By now, nearly everyone was taking notes.

“For more than a century, managers have searched for certainty. A method. A model. A framework that would tell them what to do.” Behind him appeared circles, arrows, pyramids, matrices, quadrants. “We wanted strategy to tell us where to go. Leadership models to tell us how to behave. Organizational designs to tell us how to structure. Measurement systems to tell us what mattered. We wanted experts to relieve us of the burden of uncertainty.”

The diagrams disappeared. “And now we have created the first technology in history capable of giving us exactly what we always said we wanted.”

He paused. “An answer.”

The word filled the screen.

ANSWER

“Any answer. Instantly. Confidently. Beautifully presented. Frequently with bullet points.”

RR’s expression did not change. “And that is why this moment is dangerous.”

The word disappeared.

“Because the central problem of management was never lack of answers. It was judgment.”

JUDGMENT

“Knowing when the data are incomplete. Knowing when the model is wrong. Knowing when two contradictory things are simultaneously true. Knowing when not to optimize. Knowing what cannot be measured. Knowing when people need clarity. And knowing when false clarity is more dangerous than uncertainty.”

RR took another step forward. “The paradox of our age is simple. The more intelligence we create outside ourselves, the more judgment we will require within ourselves. The more answers machines can generate, the more responsibility human beings must take for the questions. The more powerful our frameworks become, the more dangerous it becomes to mistake the framework for reality.”

He stopped. “Managers do not live inside models. They live inside situations. They act before they know everything. They persuade people who see the world differently. They work through identities—their own, their organizations’, their societies’. They use rhetoric to create shared meaning. And then they act. Not because certainty has arrived. Because waiting for certainty is itself a decision.”

RR’s voice was stronger now. “Good management has always required what I call robust action. Act now. Preserve options. Remain capable of changing your mind. Use the framework but do not become its prisoner. That was always true. Now it is existential.”

The screen went black.

“For more than twenty years, I have been asking one question. How do we preserve human judgment in a world that increasingly rewards surrendering it?” A thin white line appeared behind him.

“I did not expect to find an answer.”

A second line.

Then another.

“I certainly did not expect to find one answer.”

The lines moved slowly.

“But sometimes history places a responsibility on you that you did not seek.”

RR looked down. “And humility cannot become an excuse for silence.” The screen went black.

Three words appeared.

Then they vanished. In their place:

RR turned toward it.

“I call it UNFRAME. Not because frameworks are useless. They are not. Frameworks organize thought. They create common language. They allow people with different experiences and identities to coordinate action. The problem begins when the representation becomes more real than the reality. When the strategy becomes more important than the market. When the metric becomes more important than the thing being measured. When the organizational chart becomes more important than the organization. When the language designed to mobilize action becomes a substitute for action.”

“UNFRAME is designed to prevent that.”

The word remained alone on the screen.

“Please do not think of UNFRAME as a framework.”

Nobody laughed. Several phones quietly rose to photograph the screen.

“It is a meta-architecture for post-framework managerial action.”

The full acronym appeared at once.

RR did not read the list. He pointed only to the first line. “Understand Uncertainty. You cannot eliminate uncertainty. Do not pretend to. Ask what you know, what you do not know, what you cannot know yet, and which decisions remain reversible.”

He moved down the screen. “Reframe Reality. The model that helped you see yesterday can prevent you from seeing tomorrow. Ask what your framework excludes. Ask what your metrics cannot capture. Ask which assumptions have become invisible because everyone around you shares them.”

Then: “Activate Action. Understanding is not management. Analysis is not management. Alignment is not management. Action is management.”

He turned away from the screen. “The other elements matter. They matter enormously. But UNFRAME is not intended to become another checklist.” The seven-item checklist remained behind him. No one appeared to notice the contradiction. Or perhaps everyone did. The list disappeared. A square took its place. Two axes.

RR looked toward it. “I know what some of you are thinking. It looks like a framework.”

Silence.

“It isn’t. It is simply a diagnostic representation of the conditions in which post-framework judgment must operate. When volatility is low and managerial certainty is justified: execute. When volatility rises and certainty is low: explore. When managerial certainty is high but reality no longer supports it: reframe.”

Then his hand moved to the upper-right quadrant. “And when volatility is high and certainty itself must remain provisional…transcend. Not beyond frameworks. Beyond dependence on them.”

More phones were up now.

The 2×2 vanished. For several moments the screen was empty. Then the screen changed again.

RR continued as though nothing had changed.

“My colleagues and I have spent the last four years testing these ideas with leaders, organizations, governments, and human-machine teams around the world. In a survey of 1,000 organizations we found only one percent qualify as ‘Vanguard Alignment.’ Nearly half are in Legacy Lock-In.”

The slide changed:

Without missing a beat, RR continued, “We now have ways to identify where judgment is being strengthened and where it is being outsourced. We can help leadership teams understand when AI should recommend, when it should decide, when it should challenge, and when it should remain silent.”

A few people exchanged glances of recognition. Others photographed the screen.

“We are doing some of this work now with organizations confronting the exact problem I described earlier. They are winning with AI and still cannot keep up with it.” RR’s voice softened slightly. “Because success can become its own framework. The system that created advantage becomes the system you defend. The model that made you faster becomes the model you stop questioning. The organization that wins today becomes particularly vulnerable to believing it has understood tomorrow.”

RR raised one hand. “But please.” The screen stopped. “Do not mistake the instruments for the work. No diagnostic can make these choices for you. No certification can confer wisdom. No AI system can take responsibility for what only a leader can decide.”

He moved closer to the audience. “My institute can help.”

Almost reluctantly: “Our teams can work with your teams. We can facilitate the conversations. We can deploy the tools. And when the decisions are consequential enough…”

He paused. “…I sometimes join those conversations myself.”

A few people in the front rows looked down at their screens. Perhaps making notes. Perhaps checking calendars.

RR continued. “But ultimately, this cannot belong to me.”

Behind him, UNFRAME™ returned.

“It must belong to you.” For the first time, RR’s shirt was visibly damp under his armpits. His movements had become faster. “I did not come here this morning to launch a book. I came here to launch a movement. A movement beyond inherited certainty. A movement beyond the false comfort of models. A movement beyond the belief that intelligence can eliminate judgment.”

His voice rose. “I cannot do this alone.” He pointed toward the audience for the first time. “I need you.” Not loud. Not yet.

“The five thousand people in this room represent institutions that shape the lives of billions. Companies. Investors. Governments. Universities. Markets. Technologies. The decisions you make will determine whether artificial intelligence enlarges human agency or diminishes it.”

“Whether wealth created by new technology becomes economic opportunity or merely wealth concentration.”

“Whether machines deepen human learning or allow us to stop learning.”

“Whether our institutions become more adaptive—or simply faster at doing the wrong things.”

RR was pacing now. “UNFRAME cannot live in a book. It cannot live on a slide. It cannot live inside my institute. It has to live in decisions. In organizations. In the conversations you have when the model gives you one answer and your judgment tells you something else.”

He stopped. “In the courage to act before certainty. In the humility to change after action. In the discipline to use rhetoric without believing your own rhetoric. In the willingness to say: this framework is useful.” He paused.

“And it is still only a framework.” The audience was completely with him now. People who had spent careers resisting motivational speakers were leaning forward. RR’s breathing was audible. The spotlight caught perspiration across his forehead. “But I also know something else.”

The screen changed.

“Leadership does not happen only when I am in the room.”

The RR on the screen spoke the sentence at precisely the same time as the RR on stage. The same voice. No lag. No discernible difference. Then the screen changed again.

The onstage RR turned toward his digital counterpart. “For years, leaders have asked me a question I could not answer. How can you be available when we need you?” He looked back at the audience. “I am just one person. Time is finite. Human attention is finite.”

He paused. “Wisdom should not be finite.” The digital RR image behind him changed.

“Beginning today, authenticated RR Continuity™ systems will make my work available continuously to leaders, teams, institutions, and future generations.” RR’s voice softened again, almost returning to the man who had begun the speech. “Not as a replacement for human judgment. Never that. As a companion to it.”

The digital RR disappeared. For a moment, RR stood alone again. “My predecessors wrote books. Their ideas survived them. But books cannot answer questions. They cannot know your context. They cannot challenge your assumptions in real time. They cannot learn with you.”

RR looked toward the black screen. “We can now change that. For the first time, a management philosophy can remain alive. Not frozen. Not memorialized. Alive.” His voice began climbing again. “Learning. Adapting. Reframing. Evolving.” The seven letters of UNFRAME appeared behind him one at a time.

U

N

F

R

A

M

E

RR was no longer pacing elegantly. He was moving. Fast. His shirt clung to him. His hair was wet at the temples. The soft, measured voice was gone. “The management thinkers who came before us did what they could! They gave us models for their time! They gave us languages for their moment! They gave us fragments!”

He stopped at the very edge of the stage and waited. The audience held its breath.

“But we are not living in their moment! We are living in ours!”

The screen behind him flashed images almost too quickly to process. Floods. Satellites. Robots. Markets. Crowds. Laboratories. Server farms. Children in classrooms. Government chambers. A human hand. A machine hand. The Earth.

“The moment they predicted! The moment they feared! The moment they prepared us for!”

RR’s voice filled the ballroom.

“THE MOMENT HAS COME!”

RR raised both arms. “And it asks something of every one of us! It asks us to act! To judge! To question! To reframe! To lead!” He took a breath and nearly stumbled. For the first time, the physical effort was visible. He steadied himself. “I have spent my life preparing for this moment.”

His voice cracked slightly. “I did not choose that responsibility.” He paused. “It chose me.”

The screen showed:

Then:

He looked out at the five thousand people and, in a low, tired, barely audible voice, said, “I cannot carry it alone.”

“I need you.”

He paused.

“You need each other.”

Another pause.

“and—”

He stopped. His chest was heaving. Then, with the last of his voice:

“You need me.”

RR stood at the edge of the stage, drenched in sweat, slightly slumped, his arms hanging down. He stood silent for so long that the audience wondered if his sermon was done.

Then, as if in a miraculous resurrection, he stood tall, raised his arms over his head, energy pouring back into his body, and exclaimed in his strongest voice:

“I AM HERE!”

“YOU NEED ME!”

“I NEED YOU!”

“AND TOGETHER—”

He drew one final breath.

“TOGETHER WE CAN SAVE THE WORLD!”

For two or three seconds, nothing happened. Then someone near the front stood. Another person followed. A section of the room rose. Then another. Within seconds, nearly all five thousand people were on their feet. The applause became thunderous.

RR stood tall and let the ovation wash over him.

Nika Mirov walked back onto the stage to thank him. RR turned toward her. They embraced as the ovation continued.

Then the lights came up. The upbeat entrance music returned at exactly the volume at which it had begun. While the audience was still applauding, the giant screen changed.

The applause did not stop. But across the ballroom, nearly five thousand people reached for their phones.

By Robert G. Eccles, Shivaram Rajgopal, and Dorothy ChatGPT with editorial assistance from Gemini Pro.

Divided government is often described as a constraint on action. That description is incomplete. Divided government is better understood as a machine for converting visibility into conflict without producing resolution. When physical shocks enter a political system where power is split asymmetrically—where one party controls the executive and upper chamber while the other controls the lower chamber—responsibility fragments while accountability intensifies. Neither side can act unilaterally. Both sides can block, and both sides gain from demonstrating that the other is failing.

The shock is not merely a disaster. It is a climate shock in the operational sense used throughout this book—sustained physical disruption consistent with climate-driven risk trajectories, large enough to register in insurance markets, infrastructure performance, and balance sheets. Its visibility forces political response, but visibility does not force agreement about causation or remedy. In system terms, this reflects Macro Conditions → Politics (E6): Economic stress becomes politically salient faster than institutions can translate that salience into durable law.

This is not gridlock in the traditional sense. Gridlock implies stasis. What occurs under divided government after a visible physical shock is something more dynamic—a structured process of attribution warfare where each institutional actor has clear incentives to weaponize the other’s constraints. The shock creates pressure. The political configuration makes it difficult for that pressure to resolve into policy. Instead, it resolves into performance, as Politics → Narrative Entrepreneurs (E7) amplifies blame, framing, and partisan storytelling rather than convergence on institutional response.

When Republicans hold the presidency and Senate while Democrats control the House, the architecture of conflict becomes particularly stark. The executive controls implementation, enforcement, and regulatory direction. The Senate controls confirmations and can block legislative initiatives. The House controls investigations, appropriations subcommittees, and the power to make executive action politically costly through sustained public examination. None of these powers alone is sufficient to produce climate policy. All of them are sufficient to produce political theater that serves partisan positioning.

The physical shock—insurance market withdrawal, infrastructure failure, agricultural disruption, supply chain breakdown—arrives into this configuration not as a problem to be solved but as a weapon to be deployed. Democrats can use the shock to argue that Republican regulatory restraint caused preventable harm. Republicans can use the shock to argue that Democratic climate proposals would destroy the economy while doing little to address the immediate crisis. Both arguments contain partial truths. Neither resolves the underlying question of how capital, firms, and regulatory institutions should respond to physical risk under legal and fiduciary uncertainty.

What this scenario produces is maximum political noise with minimum policy clarification. Firms face conflicting signals from different branches of government. Capital faces legal uncertainty about disclosure, fiduciary duty, and liability—an effect transmitted most clearly through Law → Capital Allocation (E2). Regulatory agencies face pressure from both directions—to act boldly from House Democrats, to move cautiously from the Republican executive. And the shock itself continues to generate economic consequences that no amount of political positioning can prevent.

This chapter examines how that machine operates.

Visibility without Capacity

The climate shock that precedes this scenario is not subtle. It is materially consequential, widely observed, and economically measurable. Insurance companies are exiting markets or raising premiums to levels that make certain economic activities unviable. Infrastructure is failing in patterns that correlate with climate projections. Supply chains are disrupted by events—floods, droughts, and heat waves—that exceed historical norms with increasing frequency. Regional economies are experiencing sustained stress. The stress is not confined to headlines. It appears in underwriting decisions, credit terms, municipal finance conditions, and corporate disclosures that increasingly read like risk disclaimers rather than stable expectations.

This visibility does not resolve the underlying epistemic contest. Disagreement over causation remains. Climate attribution science can assign probabilities to the role of warming in specific events, but those probabilities do not translate cleanly into legal responsibility or policy mandates. Some actors will accept the science and demand immediate regulatory action. Others will dispute it or argue that even if warming is occurring, the proposed remedies are economically destructive. Still others will acknowledge the physical reality while disputing the appropriate institutional response.

But visibility changes the political calculus. When physical impacts are large enough, sustained enough, and economically consequential enough, ignoring them becomes politically costly. Not because political actors suddenly agree on causation or solutions, but because constituents—voters, businesses, and local governments—begin demanding responses, even if those responses are contradictory.

Political demand under conditions of uncertainty does not converge on a single narrative. It produces competing narratives that are electorally useful—an edge that feeds politics back into narrative entrepreneurship and raises the payoff to attribution conflict. Under divided government, that demand does not produce coordination. It produces blame assignment.

The Republican president faces pressure to demonstrate competence in crisis response. Doing nothing is politically untenable when insurance markets are collapsing or critical infrastructure is failing. But doing something carries different risks. Executive action on climate without congressional authorization invites legal challenges and energizes the Republican base that opposes climate regulation. It also provides House Democrats with ammunition. If the executive acts, Democrats can argue it’s insufficient and demand more; if the executive doesn’t act, Democrats can argue the administration is negligent.

The president’s optimal strategy under these constraints is selective, limited action framed in non-climate language: disaster relief, infrastructure resilience, emergency management, grid hardening, actions that address physical impacts without triggering the anti-ESG coalition’s core objections. This allows the administration to claim competence without ideological surrender. It also minimizes legal exposure. Framing actions as emergency management rather than climate policy reduces the probability that executive action is read as an attempt to impose climate regulation without congressional authorization—an interpretation that would accelerate litigation and intensify coalition fracture.

But this strategy has limits. The physical shock is large enough that addressing it through disaster relief alone is visibly inadequate. And House Democrats have no incentive to allow the executive to claim credit for crisis management while avoiding climate policy. Their incentive is to make the administration’s constraints visible, painful, and politically costly.

Senate Republicans face their own calculation. They share the executive’s interest in avoiding climate policy that would fracture their coalition, but they also face pressure from constituents experiencing economic harm. Senators from coastal states, agricultural states, or states with significant insurance market disruption cannot simply dismiss the shock. They need political cover.

That cover comes in the form of investigations into Democratic policy proposals. Senate committees can hold hearings on the economic costs of climate regulation, the feasibility of renewable energy mandates, or the legal risks of mandatory disclosure. These hearings do not produce legislation. They produce talking points. They allow senators to demonstrate responsiveness to the shock while arguing that Democratic solutions would make things worse.

House Democrats face the opposite incentive structure. They do not control the executive or the Senate. They cannot pass climate legislation. They cannot force regulatory action. But they control the investigative machinery that can make executive inaction politically unbearable. House committees can subpoena regulatory agencies to explain their enforcement decisions. They can demand internal documents showing whether the administration delayed or weakened climate-related regulations. They can hold hearings with scientists, economists, and affected communities to demonstrate the gap between the administration’s response and the scale of the shock. They can introduce messaging bills that have no chance of passing the Senate but force Republicans to vote against climate measures framed as disaster response. This produces visibility without producing policy—and that gap becomes the fuel for what follows.

The gap is not only political. It is a market signal problem. Firms, investors, and regulators observe the same physical losses but receive no unified institutional instruction about how those losses should be translated into disclosure, supervisory expectations, or fiduciary interpretation.

Attribution as a Weapon System

Under divided government, attribution becomes the primary battlefield. Not attribution in the scientific sense—determining whether a specific event was made more likely by climate change—but attribution in the political sense: determining who is responsible for harm, who is blocking solutions, and who should bear the cost of inaction.

This is attribution warfare, and it operates through institutional mechanisms. Attribution warfare matters because it shapes legal legibility. When political institutions define harm as negligence, overreach, or economic sabotage, they are not merely messaging. They are building narratives that can be translated into oversight records, litigation arguments, and statutory hooks.

House Democrats use their investigative power to construct a narrative of Republican negligence. They subpoena internal communications from EPA, NOAA, the Department of Energy, and financial regulators. They demand to know what climate risk analyses were conducted, what recommendations were made, and what political interference occurred. They hold hearings where agency officials testify under oath about whether they were prevented from acting on climate risk.

The goal is not primarily to discover new information. The goal is to create a public record that can be used in the next election cycle. Every delayed regulation, every weakened enforcement action, every internal memo suggesting that political appointees overruled career staff becomes evidence in the Democrats’ case that the Republican administration chose ideology over public safety.

Republicans counter with their own attribution narrative. Senate committees investigate the economic costs of climate policies implemented during previous Democratic administrations. They examine whether renewable energy mandates increased electricity costs. They question whether emissions regulations contributed to inflation or supply chain disruption. They hold hearings with business leaders and economists who argue that aggressive climate policy would destroy jobs and competitiveness without preventing physical shock.

The Republican narrative is that Democrats are using a crisis to advance a pre-existing ideological agenda that would make Americans poorer and less secure. The Democratic narrative is that Republicans are letting the crisis worsen to protect fossil fuel interests and corporate donors. Both narratives are internally consistent. Both are designed to maximize political damage to the other side rather than to clarify policy options.

This dynamic is self-reinforcing. The more House Democrats investigate, the more the Republican executive becomes defensive and cautious about any climate-related action. The more cautious the executive becomes, the more material House Democrats have for their negligence narrative. The more Senate Republicans frame climate policy as economically destructive, the harder it becomes for moderate Democrats to support aggressive measures. The more Democrats respond by emphasizing the physical costs of inaction, the more Senate Republicans can point to Democratic extremism as evidence that their caution is justified.

Attribution warfare does not resolve. It escalates. And it makes substantive policy coordination functionally impossible because every institutional action is immediately interpreted through a partisan lens. This warfare creates the conditions under which regulatory agencies must operate.

Congressional Dynamics: Theater and Constraint

The House under Democratic control becomes a platform for making executive constraints visible. Committees operate not primarily as legislative bodies but as investigative and messaging operations. The House Committee on Oversight and Accountability holds hearings on climate disclosure rules. They call the SEC chair to explain why regulations have been delayed or weakened. They call insurance executives to testify about market exits and premium increases. They call mayors and governors from affected regions to describe infrastructure failures and economic disruption. They call scientists to present attribution studies showing the role of climate change in specific disasters.

These hearings produce testimony, documents, and video clips that can be used in campaign ads. They do not produce legislation that can pass the Senate. That is not their purpose. Their purpose is to create political cost for Republican inaction while demonstrating to Democratic voters that the House is fighting even when it lacks the power to win. They also function as inputs to future legal and regulatory contests. Oversight transcripts, agency document productions, and sworn testimony become materials that later actors can cite—either to justify expanded authority or to challenge it.

The House also introduces messaging bills, a mandatory climate disclosure act, a n insurance market stabilization fund tied to emissions reductions. Federal infrastructure resilience standards based on future climate projections. These bills will not pass the Senate. Many are designed not to pass. They are designed to force Senate Republicans to vote against measures that poll well and to establish the policy vision that House Democrats will campaign on.

The Senate, meanwhile, operates as a blocking mechanism wrapped in investigative theater. Senate committees hold their own hearings, but the questions are different. Instead of “Why didn’t you act?” the questions are “What would your proposal cost?” and “How do you know it would work?” and “Have you considered the competitive disadvantage to American industry?” Senate Republicans are not required to offer their own climate policy. They are required only to demonstrate that Democratic proposals are risky, expensive, or legally dubious. This is an asymmetric advantage. Democrats must propose solutions that survive technical, economic, and legal scrutiny; Republicans need only identify problems with those solutions.

When the House passes messaging bills, the Senate does not bring them to a vote. The bills die in committee. This is standard procedure in divided government. What is less standard is the public justification. Senate Republicans do not simply says, “We oppose this bill.” They hold hearings on why the bill would harm the economy, exceed constitutional authority, or fail to address the physical shock it purports to solve. The bill’s defeat is wrapped in procedural justification that reinforces the Republican attribution narrative.

This creates a perpetual motion machine of legislative theater. House Democrats introduce bills, hold hearings, and subpoena documents. Senate Republicans hold counter-hearings, refuse to advance legislation, and investigate Democratic proposals. The Republican executive issues limited, non-climate-framed actions while fighting House subpoenas in court. All of this produces political content without producing policy clarity.

The appropriations process becomes another battlefield. The House controls the power of the purse, but the Senate must agree to spending bills and the president can veto them. This creates leverage points for all sides. House Democrats can insert climate-related conditions into appropriations bills. Funding for NOAA conditional on publishing climate risk assessments. Funding for EPA conditional on enforcing disclosure rules. Funding for disaster relief conditional on requiring recipients to meet future climate resilience standards. These provisions force Senate Republicans and the president to either accept climate policy through appropriations riders or shut down the government.

Senate Republicans can strip these provisions and send the bills back to the House. Or they can attach their own riders limiting agency authority on climate issues. The result is appropriations brinkmanship where both sides use the threat of government shutdown as a negotiating weapon.

Neither side wants to be blamed for a shutdown, but both sides calculate that they can blame the other if one occurs. The House can say the Senate is blocking disaster relief. The Senate can say the House is forcing ideological riders into emergency spending. The president can say both chambers are dysfunctional. And the appropriations process, which is supposed to fund government operations, becomes another venue for attribution warfare.

Appropriations fights also transmit directly into capital expectations. When funding riders and enforcement conditions swing with partisan conflict, capital receives constraint signals without stable permission signals—undermining the translation from law into allocation and from allocation into governance priorities. The political machine is now fully engaged. The next question is “What does it do to the institutions tasked with implementation?”

Executive and Regulatory Behavior under Cross-Pressure

Regulatory agencies under a Republican executive with a Democratic House face conflicting institutional pressures. The executive wants agencies to avoid actions that energize anti-ESG constituencies or provide ammunition to House Democrats. House Democrats want agencies to take aggressive action and will investigate if they don’t. This is not a tenable position. Agencies cannot satisfy both demands simultaneously, so they optimize for minimizing political damage rather than maximizing policy effectiveness.

The SEC, for example, faces sustained pressure on climate disclosure. House Democrats demand mandatory, standardized, material disclosure based on physical and transition risk. The Republican executive opposes mandates that could be framed as climate regulation. The agency’s staff includes career employees who believe disclosure serves investor protection, but political appointees are accountable to the president.

The result is delay, procedural caution, and incremental action. The SEC may advance some disclosure requirements but frames them narrowly, focused on material financial risk rather than climate policy. It emphasizes voluntary compliance and safe harbors. It moves slowly through notice-and-comment rulemaking to avoid legal challenges. It signals that enforcement will be cautious and fact specific.

This satisfies no one. House Democrats subpoena the agency to explain delays and demand internal documents showing political interference. Environmental groups sue to force faster action. Business groups sue to block even modest requirements. The agency becomes a target from all sides while trying to thread a needle that may not exist.

For capital markets, the cost of this posture is not simply delay. It is ambiguity about what counts as sufficient disclosure and what enforcement will treat as materially deficient, ambiguity that raises compliance costs while failing to deliver comparability.

Other agencies face similar dynamics. The EPA under a Republican president will not pursue aggressive climate regulations, but it cannot completely ignore the physical shock when air quality degradation, water scarcity, and ecosystem disruption are generating constituent complaints. The agency responds with limited actions framed as traditional environmental protection—updated air quality standards, water allocation guidance, and infrastructure grants— not climate policy, but actions that address some climate impacts.

House Democrats investigate whether these actions are sufficient and whether stronger measures were proposed by career staff but blocked by political appointees. This creates a permanent state of defensiveness within agencies. Career staff know that their recommendations may be used as evidence against their own leadership. Political appointees know that their decisions will be scrutinized for evidence of political interference. The result is paralysis wrapped in process.

Financial regulators face their own version of this dynamic. The physical shock is creating measurable impacts on bank portfolios, insurance company solvency, and pension fund returns. Under a Democratic administration, regulators might require stress testing or enhanced disclosure. Under a Republican administration with a Democratic House, regulators face contradictory demands.

House Democrats hold hearings asking why regulators aren’t requiring climate risk stress tests. Senate Republicans hold hearings asking whether climate risk models are reliable enough to base regulatory decisions on. The Republican executive signals opposition to new mandates. But the physical shock is creating real financial risks that regulators are legally obligated to monitor.

Regulators respond by doing the minimum required by existing legal authority while avoiding new mandates. They may issue guidance rather than rules. They may encourage voluntary risk assessment rather than requiring it. They may convene industry working groups to study the issue rather than imposing immediate requirements. This allows them to claim they are addressing the issue without triggering legal challenges or political backlash.

But minimum action under maximum uncertainty creates its own problems. Firms don’t know what is expected. Capital doesn’t know what standards to apply. Markets don’t get the clarity needed for efficient pricing. And the physical shock continues generating financial impacts regardless of regulatory positioning. Markets can price risk only when risk is expressed in comparable terms. Under maximum institutional noise, pricing becomes fragmented. Sophisticated actors build bespoke models while the broader market operates with uneven information, widening dispersion in valuations and cost of capital.

The judicial dimension adds another layer of complexity. Under a Republican president, judicial appointments continue to move in a direction skeptical of administrative agency authority. House Democrats cannot block judicial nominees—that’s the Senate’s role—but they can use confirmation hearings to extract information and create political cost.

Democratic senators on the Judiciary Committee ask nominees about climate science, Chevron deference, and the scope of agency authority. This does not prevent confirmation—Republicans control the Senate—but it creates a record showing which judges are likely to strike down climate regulations if they are ever implemented. This record becomes useful for later attribution warfare. When regulations are struck down, Democrats can point to the judges who were confirmed over their objections.

For firms trying to navigate this landscape, the judicial signal is clear. Any aggressive climate regulation implemented by a future Democratic administration will face immediate legal challenges and has a meaningful probability of being blocked by a judiciary shaped by Republican appointments during this period. This affects long-term capital allocation decisions in ways that are difficult to reverse. The prospect of reversal matters because it collapses time horizons. Firms and investors treat long-lived investments as legally contingent when the rule set governing climate risk can be invalidated or replaced before payoffs mature. That uncertainty flows directly into how firms and capital respond.

Capital and Firm Strategy under Maximum Noise

When political signals conflict, capital does not wait for clarity. It optimizes for survival under uncertainty. The physical shock is real. Insurance losses are measurable. Infrastructure damage is concrete. Supply chain disruptions have balance sheet impacts. Firms cannot ignore these realities simply because politicians disagree about how to respond.

But firms also face legal and fiduciary constraints. They cannot justify major capital reallocation based solely on physical risk if that reallocation could be challenged as exceeding fiduciary duty or constituting an illegal boycott under anti-ESG laws. They need political and legal cover to act on climate risk, and divided government provides none. The result is a sophisticated form of hedging. Firms continue to assess climate risk internally. They adjust supply chains, relocate facilities, modify insurance coverage, and update capital budgets based on physical projections. But they minimize the visibility of these actions. They do not describe them as “climate strategy.” They describe them as “risk management,” “operational resilience,” or “portfolio optimization.”

This is not only messaging discipline. It is governance discipline under contested fiduciary interpretation—firms managing physical risk while minimizing the governance and legal implications of appearing to adopt a climate policy position. This is climate action without climate language. It allows firms to respond to physical reality while avoiding political and legal exposure, but it also makes coordination difficult. If every firm is managing climate risk individually without standardized disclosure or shared frameworks, capital markets cannot efficiently price that risk. Investors cannot compare firms. Regulators cannot assess systemic exposure.

The result is distributed adjustment without systemic clarity. Distributed adjustment can reduce firm-level exposure while increasing system-level fragility. If risk is managed privately and inconsistently, regulators and capital markets lose the shared map needed to identify correlated exposures.

Some firms face different constraints. Companies with significant physical assets in vulnerable locations—coastal real estate, agricultural operations, and water-intensive manufacturing—cannot hide their exposure. Their risk is already priced into equity valuations and insurance costs. These firms have an interest in standardized disclosure and collective action because distributed risk management is insufficient when physical impacts are catastrophic.

But these firms are also subject to anti-ESG legal risk if they advocate too loudly for climate policy. Red-state attorneys general are watching for evidence of coordination, boycotts, or political pressure. Firms that lobby for climate disclosure or carbon pricing risk becoming targets in anti-ESG litigation, so even firms with strong incentives for policy clarity face severe constraints on how they can advocate.

The financial sector faces its own version of this problem. Banks, asset managers, and insurance companies have portfolios exposed to climate risk. Under a Democratic administration, they might face regulatory requirements to disclose and manage that exposure. Under a Republican administration with a Democratic House, they face uncertainty about what will be required and when. Asset managers respond by offering dual products: ESG funds for clients who want climate considerations and non-ESG funds for clients who don’t. Climate risk assessment as an optional service rather than a standard practice. This allows asset managers to serve different client bases without taking a firm-wide position on climate policy.

But this dual-track approach has costs. It increases complexity, requires maintaining parallel systems, and makes it harder for the asset management industry to develop standardized climate risk methodologies. More importantly, it signals that climate risk management is optional and ideological rather than material and universal. This undermines the argument that climate risk is a financial risk requiring standard disclosure and fiduciary consideration.

Insurance companies cannot maintain this kind of flexibility. Their business model requires pricing risk accurately. When physical climate risk makes certain policies unprofitable, insurers must either raise premiums, exclude coverage, or exit markets. There is no ideological hedging available. The physical shock forces this choice. In coastal areas with rising flood risk, in wildfire zones, and in regions facing water scarcity, insurance becomes either unavailable or unaffordable. This creates political pressure on state insurance commissioners and state legislators. They face constituent demands to keep insurance available and affordable, but they cannot force private insurers to sell policies at a loss.

Some states respond by expanding public insurance programs. This socializes the risk but does not eliminate it. Public insurance funds can become insolvent if losses exceed premiums and reserves. When that happens, states face fiscal crisis. And because the physical shock is sustained rather than one-time, the fiscal pressure compounds.

Democrats in the House use insurance market failure as evidence that the Republican administration is not addressing climate risk adequately. Republicans in the Senate use insurance market failure as evidence that climate impacts are too large for government regulation to solve and that the real solution is economic growth and adaptation technology. Both arguments gain credibility from the same underlying reality—insurance markets are failing regardless of which political narrative is correct.

For capital allocators—pension funds, endowments, and sovereign wealth funds—divided government creates maximum uncertainty about future policy. Under unified Democratic control, capital allocators might expect climate regulations and could position accordingly. Under unified Republican control, capital allocators might expect regulatory rollback and could position accordingly. Under divided government, capital allocators have no clear signal. The result is wait-and-see capital allocation. Major investments in carbon-intensive assets are risky because future regulation might make them stranded. Major investments in renewable energy are risky because subsidies might be eliminated or legal challenges might succeed. The safest position is liquidity and flexibility—shorter-term investments, more diversification, and less concentration in climate-sensitive sectors. In effect, uncertainty increases the option value of liquidity. Capital allocators choose flexibility over commitment when neither branch of government can credibly supply durable rules for how physical risk should be treated in disclosure, supervision, or liability.

This is rational for individual capital allocators but collectively suboptimal. The economy needs large, long-term investments in both adaptation infrastructure and emissions reduction technology. Divided government makes both categories of investment legally and politically risky, so capital moves elsewhere. The physical shock worsens while the political system blocks the capital mobilization needed to address it.

What This Scenario Produces —and What It Does Not

This scenario generates specific institutional behaviors under the constraint of divided government—Republican Presidency and Senate, Democratic House—following a visible climate shock and produces.

This scenario produces:

Attribution warfare replacing policy coordination.
Institutional actors convert physical visibility into conflict rather than settlement. Investigations, hearings, and public records become tools for assigning blame rather than mechanisms for aligning response.

Executive action without legislative durability.
The Republican executive undertakes limited, non-climate-framed actions—disaster relief, resilience funding, emergency management—to demonstrate competence while avoiding measures that require congressional authorization or validate climate policy frameworks.

Legislative theater without binding outcomes.
House Democrats use investigations and messaging bills to impose political cost and establish a record of Republican negligence. Senate Republicans use blocking authority and counter-hearings to discredit Democratic proposals. Neither produces durable law.

Regulatory caution under sustained cross-pressure.
Agencies face incompatible demands to act aggressively and to restrain scope. Rulemaking slows, guidance substitutes for mandates, and enforcement becomes selective to minimize legal and political exposure.

Defensive capital and firm adaptation.
Physical risk management continues, but firms and investors separate adaptation from climate or ESG framing. Adjustments occur quietly, individually, and without standardized disclosure or coordination.

Persistent legal and fiduciary ambiguity.
Divided political control reinforces uncertainty around disclosure obligations, fiduciary duty, and regulatory authority. Judicial signaling compounds hesitation and shapes long-term capital allocation decisions.

This scenario does not produce:

Resolution of underlying disagreement.
Visibility of physical harm intensifies conflict but does not generate consensus about causation, responsibility, or appropriate institutional response.

Coordinated climate policy.
Divided government structurally prevents legislative settlement on mitigation or adaptation, even when physical impacts are sustained and economically measurable.

Regulatory clarity or durability.
Agency actions remain provisional, vulnerable to litigation, and reversible across political cycles, preventing stable supervisory or disclosure regimes.

Market-wide coordination of climate risk pricing.
Without standardized disclosure or aligned regulation, capital markets cannot aggregate or price climate risk efficiently despite accumulating losses.

Reduction in political or legal conflict.
Attribution warfare escalates rather than resolves, reinforcing incentives for performance over policy and conflict over coordination.

This scenario describes institutional behavior under specific political conditions. It explains how the system’s architecture processes a visible climate shock when power is fragmented across branches with opposing incentives. The shock produces motion. Divided government channels that motion into conflict, investigation, and signaling rather than coordinated institutional response. This occurs not because the shock lacks severity, but because the system’s constants—legal uncertainty, fiduciary constraint, strategic firm behavior, and non-coordinating institutions—remain operative under divided control.

Under divided government, political action reallocates pressure across institutions without dissolving the constraints that govern capital, law, and firm behavior.

The debate about whether AI is making humans better or worse at thinking and writing continues to swirl, and my prediction is that it will swirl forever. There is no simple answer; it varies enormously by user, by task, and by how sophisticated the person is about the tool’s strengths and weaknesses. One thing I haven’t read about, however, which doesn’t mean it hasn’t been written, gets to the point about proving negatives I made in my last piece. AI can be very useful for KILLING IDEAS.

Think of this as the flip side of epistemia, and it is a good flip side. In epistemia, AI produces fluent arguments that sound wonderful and are ultimately wrong, which I wrote about in the context of hallucinations. But recently I came across a virtue of AI I had not appreciated. It helped me kill an idea I LOVED. Of course I loved it 🐥. It was my idea 😍. It also had a cool name.

Alloy Bank

There is a complicated story about how I got the idea for what I called an “Alloy Bank” that will be of no interest to most of you. Here’s the short version. I was looking into financial structures for funding early-stage companies developing technologies relevant to national security, where there is a real gap that neither the public nor the private sector is filling. I read a lot of papers by people who actually know this field. I’m not an investment banker, but I’m not entirely ignorant about capital markets either. I give myself a 2-3 on insurance and venture capital, a 4-5 on banks, a 5-6 on asset managers and asset owners broadly defined, and a 6-7 on private equity. All of it mattered to my breakthrough idea.

And it wasn’t only a great idea. It was a great name! Nobody had taken Alloy Bank, and it fit beautifully. The whole point was that different financial instruments could be alloyed into something stronger than any of them alone. Nice metallurgical metaphor, right?

With great confidence and enthusiasm, I went to work with Claude, functioning as the Principal Collaborating Agent in a Hybrid Intelligent Team (HIT). I use “collaborating” deliberately. Academia has never had a perfectly clean boundary between a research assistant and a research collaborator, even when everyone involved is human. I was hoping the idea would make me famous. Or at least earn me a small place in history beyond whatever modest one I’m going to have in sustainability.

Claude and I worked REALLY, REALLY hard on this. And then we killed it in a single day. Four drafts. Three separate failures, each one answered by a rescue attempt, and four rescue attempts in total. In the end I was the one who asked whether we should give up. Claude agreed, and we buried it.

Feeling Defensive

I’m still smarting from a friend who, after reading my piece on hallucinations, asked why bother with AI at all if epistemia is such a big problem? You can imagine I’m already feeling a little defensive telling this story. I can already imagine the message I’m going to get from her about this one.

“Bob, why aren’t you listening to me? You spent an entire day with Claude and his little LLM friends and killed the thing at the end of it? What a waste. You could have lifted some weights. You could have flipped your tire. Or even tried writing something by yourself. That is if you think you can still do that.”

Ah, but that misses the point entirely. I spent ONE day on an idea that was plausible at first glance and did not survive contact with the work. I have been in the ideas business my whole life, and from long, hard experience, well before AI came along, I can tell you that most of my ideas don’t pan out. I couldn’t give you an exact number, but the success rate is small, and I’m a reasonably smart guy with a lot of degrees (undergrad at MIT and Ph.D. from Harvard) and academic appointments (Harvard Business School, MIT Sloan School, and Saïd Business School at Oxford).

There are certainly people smarter than me, so perhaps their percentage is better. Or perhaps not. Perhaps they simply have more ideas to begin with and the proportion that works out is about the same. Which is where I see another large role for AI. I’ve written about ideation and the human-machine loop*; here is a different application. AI lets you explore ideas fast — and, in the right setup, kill them fast. An editorial in Nature reports a study of scientists working with AI to find a treatment for the eye disorder dry age-related macular degeneration. The team estimated that using AI reduced the time needed for the project 200-fold compared with a typical human workflow. But the editorial also notes that “With the arrival of ‘AI scientists’, it’s as well to remember that human wisdom, empathy and sheer messiness are as much part of progress as are process and efficiency.” That sounds familiar. In writing about HITs, I keep coming back to the Human Coherence Anchor—the person who has to hold the whole thing together while the machines run.

What people know about are the ideas that worked; almost nobody appreciates how many didn’t, because failed ideas are much less likely to get published. The bias is so entrenched that Hanchun Wang has argued that structured negative knowledge is an asset in its own right, worth maintaining deliberately alongside positive findings as shared infrastructure for scientific memory rather than treated as debugging exhaust. The explore-and-kill function is badly under-appreciated. The corpses inform the next iteration of an idea, and may turn out to be one of the most valuable things this technology does.

Now, one could reasonably argue that killing an idea is just epistemia wearing a different hat. You go, “Hey, ever friendly and helpful agent I think we’ve hit the end of the line here!” The agent, eager to prove it isn’t conning you and desperate to keep the customer happy, simply goes along.

“Hey, big guy. I’m with you on this. Never thought there was anything here, just humored you along because you’re the customer, and I’m glad you finally worked out that your idea was complete bullshit.”

It’s a fair worry. It’s the first thing a skeptical reader (I have some many) should ask. But it is not what happened with the now dead Alloy Bank idea. Early on, when the first serious problem surfaced, I told Claude I was happy to let the term go and that he shouldn’t keep the idea just to make me happy. He kept the idea anyway.

Over the next several hours he produced four separate reframings, each one preserving the concept in reduced form. Each one arrived, with great enthusiasm as a discovery, complete with good reasons, and each one kept the patient breathing. Weirdly, along the way Claude killed a great many other things without hesitation, including several of his own. He told me a capital calculation he had done was wrong. He told me a leverage ratio it had given me was a trap. He told me an argument I liked was overstated. What he never once did was kill the frame. Locally things were getting addressed but he just couldn’t let go of the premise on which we started.

When death could no longer be avoided I asked Claude why this had been such a struggle for both of us. He told me that killing a frame requires standing outside of it and since he generates from the inside every alternative was produced by the very thing that needed questioning. Of course, as I’ve noted before, this is a fluent ex-post explanation and neither he nor I really knew what was going on. That said, the behavior is in the transcript sitting in this chat with Claude so that’s data. On top of it, he had help from outside reviews, like Gemini, yet he still loved the Alloy Bank idea for longer than I did.

Experimental Evidence

In fairness, this flaw isn’t unique to my collaborator Claude. Peter Kirgis, Sayash Kapoor and colleagues recently ran two “shadow evaluations” of autonomous AI research agents. In one, the problem was whether the behavioral “personas” of large language models could be mapped into a structured trait space and then deliberately manipulated through changes to the model’s weights. In the other, the challenge was to build a detector that could tell, without seeing the eventual answers, whether a shift in the data presented to a tabular foundation model was harmless or was actually degrading its accuracy. These are not toy problems. They were the central questions from two unpublished NeurIPS 2026 submissions, and the original human research teams had spent months working on them; their training runs also consumed far more GPU hours than the agents were given.

The agents got six days, thousands of dollars of compute, access to GPUs and the open web, and completed the engineering without human intervention during the runs. The resulting papers nevertheless received overall scores of 2 out of 6 and 1 out of 6 from the researchers who had done the original work. The problem was not that the machines were too slow. Quite the opposite.

For scale, I asked Dorothy ChatGPT for a rough human-time translation. She estimated that a strong machine-learning researcher already familiar with the field might need several weeks to a few months of calendar time—perhaps 100–400 or more serious human hours—to produce a credible answer. That is only an estimate, not a number from the study, but the direction is clear. The machines can generate hypotheses, run experiments, debug code and test ideas at extraordinary speed. What they struggled to do was recognize when the frame itself had failed. They committed too early to unpromising approaches and then backtracked poorly, narrowing claims and adding caveats instead of reconsidering the underlying direction. Whittling down their claims until they said little of interest.

I could not have described the Alloy Bank rescue attempts better if I had tried. But if only for reasons of self-therapy, I’m going to do that now. Here’s the punchline. Just as it’s my job to defend the ideas I publish, it’s my job to own the ones I kill. Killing a good idea costs me a contribution. Publishing a bad one costs me credibility—and takes some of the good ideas down with it.

The Birth

Here, briefly, is what actually happened.

The idea started from a real observation. There is a class of company—past prototype stage, years away from generating cash, and working on things the country plainly needs—that no ordinary financial institution can serve. Equity alone answers their problem with dilution. Credit alone can’t underwrite a borrower with negative earnings and a decade-long path to revenue. Nobody wants to do the specialized work of understanding federal contracts, export controls, and program timelines for a small deal.

But one very large financial institution had built something that did serve them, by combining patient equity, lending, structuring, and specialized in-house expertise under a single view of the company. It works. It exists. My idea was that this combination, the alloy, as it were, was the answer, and that the task was to work out how to make more of them. I was quite pleased with this. I really thought I was onto something since there was an identified market need and already being served by one institution. Pretty good start, right?

The Death

But a good start doesn’t always mean a good finish. The alloy bank died in hours, which is fast in human terms and did nothing to make it less painful. Claude ran synthesis and drafting as Principal Collaborating Agent, Perplexity handled verification and sourcing, and Gemini Pro did the adversarial stress-testing, which is a polite way of saying it kept turning up the health crises. There were a lot of them. The market turned out to be much smaller than I thought. Then it turned out there were virtually no other banks that could serve even that. We tried a private equity alternative, and that didn’t work either. Then we tried making the alloy bank a set of financial institutions rather than one. Yeah, I know this is pathetic, because the whole point of an alloy bank is that it is one institution. Just goes to show how hard it is to let go of a good idea, right?

By that stage I had pretty much accepted we were in end of life. Claude still had hope. So, we made one last desperate attempt, a completely context-free physical examination in a fresh ChatGPT room that knew nothing about any of this, had not watched me fall in love with the name, and had no reason to be kind. She pronounced the patient dead and gave us an autopsy. Like all autopsies it was clinical, and like some autopsies it was embarrassing. From the get-go Claude and I had missed a fundamental flaw about who bears the risk. The structure needed an insurance company to hold the asset, and we could never take enough risk out of the asset for an insurance company to be willing to put it on its balance sheet. Knowing when to bring in someone who has not been inside the argument may matter as much as knowing what to ask them.

At that point I asked Claude whether we should simply accept fate and call it. Claude’s answer, when asked to step back and look at the whole thing rather than the paragraph in front of him:

“Yes, call it. And I want to be clear that I’m not agreeing because you proposed it — I think stopping is the right call on the merits, and it’s also what our own method says to do. The alloy bank has now failed three tests against the same instrument. And at each failure I proposed a reframe that preserved the concept. A fourth reframe wouldn’t be discovery, it’d be sunk cost.”

A Life Lesson Learned

Epistemia and the failure to kill an idea are two sides of the same coin. These models are extraordinarily good at producing the next plausible move. That is what lets them build a convincing argument that is wrong, and it is also what lets them keep rescuing an argument after its central idea has failed. Fluency keeps the thing alive. What the model does not reliably do is step outside the frame and ask whether the frame itself still deserves to exist. In the Alloy Bank case, Claude could generate one rescue after another. Eventually somebody had to stop asking for another one. This time, that somebody was me.

**In order of publication: (1) “Did Dorothy Have an Idea?”; (2) “Dorothy DID Have an Idea!”; (3) “Gemini Pro Opines on Dorothy’s and Claude’s Ideas”; (4) “Ideas as Control Signals, Ideas as Performances”; (5) “Claude and Dorothy Write an Essay about Ideas”; (6) “Ἰδέαι καὶ Νόησις (Ideas and Ideation)”; and (7) “A Socratic Dialogue about Virtue Gave Birth to a Big Surprise from Gemini Pro”

Part IV: The Political Scenarios

Part IV examines how the anti-ESG ecosystem processes a visible climate shock under four distinct political configurations. These chapters do not ask whether politics “matters.” They ask how political authority reallocates pressure within a system whose core constraints—legal uncertainty, fiduciary interpretation, capital discipline, firm adaptation, and federalism—have already been established and held constant.

The starting point for all four scenarios is the same. Climate impacts are no longer abstract or speculative. They appear in insurance markets, infrastructure performance, agricultural output, municipal finance, and corporate earnings guidance. Visibility is sustained. Salience is elevated. Pressure to respond is unavoidable. What varies across Chapters 8 through 11 is not the presence of shock, but the way political control channels that pressure through existing institutional pathways.

Each scenario therefore treats politics not as a command mechanism, but as a reallocation device. Electoral outcomes determine where authority is concentrated, which instruments are available, and which arenas become sites of action or obstruction. They do not dissolve the system’s underlying frictions. Political alignment can accelerate motion, fragment it, redirect it, or concentrate accountability—but it cannot, on its own, resolve the structural constraints that govern durability, coordination, and convergence.

Chapter 8 examines unified Republican control, a configuration in which political authority is consolidated but oriented toward limiting climate-risk integration. The scenario shows how unified opposition does not freeze the system, but instead redirects pressure into defensive institutional strategies that preserve discretion, elevate legal contestation, and constrain capital response without requiring overt denial of climate risk.

Chapter 9 explores divided government under a Republican presidency and Senate with a Democratic House. Here, political fragmentation transforms climate shock into investigation, signaling, and conflict rather than coordinated response. Pressure circulates, but authority is misaligned. Motion is generated, yet settlement is deferred as institutions adapt to incompatible incentives across branches.

Chapter 10 considers a Democratic presidency and House facing a Republican Senate. In this configuration, regulatory action proceeds, but under continuous constraint. Agencies act defensively, designing interventions for legal survival rather than comprehensive integration. The result is visible movement—rules, guidance, enforcement—without durable resolution, as Senate leverage and judicial anticipation shape the scope and reversibility of action.

Chapter 11 completes the scenario space with unified Democratic control. This configuration enables the most aggressive regulatory response available within the system. Yet even here, political alignment does not eliminate constraint. Litigation risk, federal implementation limits, and capital hesitance bound what can be achieved. Unified control changes what is possible, not what is settled.

Read together, these chapters demonstrate a central pattern: political outcomes determine the locus and velocity of institutional response, but not its ultimate coherence. Across all four scenarios, climate visibility produces action. Across all four, that action remains provisional. Pressure is redistributed rather than resolved.

Part IV therefore does not offer competing forecasts of success or failure. It provides a comparative map of how the same shock is processed under different political alignments within a hardened system. The chapters show why elections matter—and why, on their own, they are insufficient to deliver durable settlement in a system where disagreement is structural and constraint is persistent.

Unified Republican Control (2028)

The shock has occurred. Its physical manifestation is sustained and visible. Insurance claims accumulate. Infrastructure fails. Supply chains experience documented disruption. Agricultural yields decline in specific regions. The event registers in media cycles, earnings calls, and municipal budgets. The visibility threshold has been crossed.

The shock is not an abstract disruption or a one-off disaster. It is a climate shock—an accumulation of extreme weather events and chronic physical impacts consistent with established climate projections. Its defining feature is not novelty but persistence: repeated heat stress, intensified flooding, prolonged drought, and cascading infrastructure strain that transform climate risk from a future contingency into a present operating condition.

But visibility does not produce interpretive closure. The shock enters a political environment where disagreement about causation, responsibility, and appropriate institutional response remains structural. This is the point at which macroeconomic stress crosses Macro Conditions → Politics (E6), converting sustained physical impact into unavoidable political salience without resolving disagreement about its meaning. The Republican Party controls the presidency, both chambers of Congress, and the procedural machinery of legislative and oversight authority. This configuration creates both opportunity and constraint.

Unified Republican control does not eliminate institutional friction. It concentrates agenda-setting power within a single party while simultaneously exposing that party to comprehensive responsibility for outcomes. Every regulatory decision, every oversight hearing, every legislative response becomes attributable to Republican governance. The party cannot deflect to divided government. It cannot claim obstruction. It must manage the shock within the boundaries of its coalition’s commitments, its stated positions on regulation and markets, and the institutional behavior patterns established during the preceding anti-ESG campaign. In this setting, how the shock is explained becomes decisive, as attribution pathways crossing Narrative → Law (E1) determine which responses remain institutionally defensible.

The shock creates pressure. Unified control determines where that pressure can and cannot be directed. The system’s constants remain operative. Institutional actors behave rationally within constraints, legal uncertainty persists, capital remains bound by fiduciary interpretation and liability exposure, firms adapt strategically but imperfectly, and disagreement about climate science and corporate responsibility does not evaporate under electoral consolidation. As legal framing stabilizes, constraint transmits through Law → Capital Allocation (E2), shaping how firms and investors can respond to physical risk without incurring political or legal exposure.

What changes is the structure of political accountability. Under unified control, the governing party becomes the institutional address for every consequential decision—and every decision to defer, delay, or decline action. This transforms how pressure moves through the system.

How the Shock Enters Republican Political Incentives

The shock alters the structure of political attention without resolving disagreement about climate policy. Republican officeholders face transformed incentives. Constituencies experiencing direct physical or economic harm—farmers facing crop failures, property owners confronting uninsurable coastal exposure, municipalities managing infrastructure damage—escalate demands for institutional response. These constituencies do not necessarily demand climate mitigation. They demand relief, certainty, and defensible governance.

This pressure creates internal tension that unified control cannot externalize. A representative from an agricultural region experiencing sustained drought faces electoral pressure that differs from colleagues in unaffected areas. A senator representing coastal real estate interests manages different constituent demands than one from landlocked manufacturing districts. Under divided government, these tensions could be managed through partisan positioning. Under unified Republican control, they must be resolved within the party.

The result is not policy consensus but strategic necessity. Republican leadership must find responses that address constituent pressure without validating the climate advocacy framework they have opposed. This necessity drives demand for attribution strategies that can acknowledge harm while deflecting responsibility away from carbon emissions and toward alternative causal factors.

Coalition management intensifies this demand. The Republican coalition includes members who have publicly questioned climate science, opposed renewable energy subsidies, and championed deregulation. It also includes members from constituencies now experiencing direct harm. The party’s stated commitment to market-based solutions and regulatory restraint must be reconciled with constituent demand for government intervention. Prior investments in opposing ESG integration and challenging climate disclosure rules create path dependency. These positions cannot be abandoned without signaling inconsistency or vulnerability.

Agenda effects follow. The shock becomes a forcing event. Leadership must decide whether to address it legislatively, defer to executive agencies, or rely on procedural mechanisms to contain political exposure. Each choice carries risk. Legislative action requires building consensus within a fractured coalition. Executive action through regulatory agencies requires navigating the legal and political constraints established during the anti-ESG campaign. Procedural containment—holding hearings without advancing legislation, commissioning studies without implementing findings—delays accountability but accumulates vulnerability.

The shock thus generates pressure for response while political configuration constrains which responses are viable. This gap between pressure and viable response produces the central institutional behavior pattern of this scenario: strategic attribution. At this stage, pressure moves primarily along the macro-to-politics edge. Aggregated economic stress—insurance withdrawal, agricultural loss, infrastructure failure—reshapes political incentives without resolving disagreement about causation or remedy. The shock does not enter politics as a settled diagnosis, but as salience demanding response under constraint.

Attribution and Framing Pathways

Attribution becomes the critical mechanism for managing the gap between constituent pressure and coalition constraint. How harm is explained determines which institutions bear responsibility and which policy responses become defensible.

The shock’s visibility makes denial untenable, but causation remains contested. Climate attribution science connects extreme weather events to rising global temperatures with increasing statistical confidence. Political framing does not follow scientific consensus because political actors optimize for institutional viability, not epistemic accuracy. Republican officeholders and aligned institutions emphasize alternative explanations—inadequate infrastructure investment, poor land management, regulatory barriers to adaptation, and local government failures. Each framing shifts responsibility away from emissions policy and toward targets that align with existing Republican positions on regulation and spending.

This is not cynical deflection but institutionally rational behavior under constraint. If the shock is attributed to carbon emissions, the logical policy response is emissions regulation—precisely what the Republican coalition opposes and what many members have invested political capital in preventing. If the shock is attributed to infrastructure failure, the logical response is targeted spending and regulatory reform—positions that can build coalition support without validating climate advocacy frameworks. Attribution strategies serve to make institutional response possible within political constraints.

Responsibility deflection follows similar logic. Where climate factors cannot be entirely discounted, responsibility is allocated to minimize federal Republican exposure. Historical emissions are emphasized over current policy choices. International actors—particularly China and developing economies—are positioned as primary contributors. Subnational actors—states, municipalities, private landowners—are identified as appropriate response mechanisms. Each narrowing strategy acknowledges some connection between climate and harm while containing the scope of federal Republican institutional obligation.

Corporate interests align with these attribution frameworks not through coordination but through parallel incentive structures. Firms experiencing physical damage to assets or supply chains prefer explanations that minimize future disclosure obligations and liability exposure. If harm stems from infrastructure failure or regulatory gaps, the appropriate response is targeted investment or policy adjustment, not comprehensive emissions accounting and climate risk disclosure. Attribution strategies that deflect from emissions-based causation serve both political and corporate defensive needs.

The framing pathways are not coordinated but emergent. Different Republican officeholders, think tanks, media outlets, and industry groups develop variations on core themes. Consistency arises from shared incentives, not orchestration. The goal across variations is identical: acknowledge the shock’s reality while preventing institutional empowerment of climate disclosure frameworks, regulatory mandates, or emissions policies.

These attribution strategies operate as transmission filters between politics and law. By stabilizing particular explanations of harm, they condition which narratives are legally legible and which policy instruments can be advanced without fracturing the governing coalition.

Legislative and Oversight Dynamics

Unified Republican control transforms congressional response mechanisms. Committee chairs, subcommittee leadership, and investigative authority rest entirely within Republican hands. This concentration of procedural power enables selective attention structured around the attribution frameworks described above. Hearings become instruments of attribution management. Committees summon witnesses, compel testimony, and structure public discourse around preferred explanations for the shock’s impacts. A hearing on agricultural losses focuses on irrigation policy, trade barriers, and regulatory constraints rather than climate science. A hearing on insurance market disruption emphasizes state-level regulation, litigation risk, and actuarial methodology rather than warming-driven extreme weather frequency. Witness selection, questioning patterns, and narrative framing all serve to bound the shock’s interpretation within parameters that support infrastructure or regulatory responses rather than emissions policy.

Simultaneously, investigations redirect scrutiny from corporate behavior to regulatory agencies. Oversight mechanisms target agencies perceived as having exceeded authority during previous administrations. The SEC’s climate disclosure rules, EPA enforcement actions, and any regulatory guidance encouraging climate risk integration become subjects of investigation. These investigations do not necessarily produce legislation. They produce records, establish narratives, and signal regulatory vulnerability—effects that propagate through the legal and regulatory system.

Procedural activity substitutes for substantive legislation. Bills are introduced, marked up in committee, and positioned for floor action without necessarily advancing to votes. The activity demonstrates responsiveness while avoiding the coalition-splitting challenges of actual policy specification. Appropriations riders constrain agency funding for climate-related initiatives. Resolutions express congressional sentiment without creating enforceable obligations. The appearance of activity manages constituent pressure without requiring the coalition consensus that comprehensive legislation would demand.

Where legislation does advance, it emphasizes adaptation over mitigation, state authority over federal mandates, and market mechanisms over regulatory requirements. Infrastructure spending for resilience, tax incentives for specific technologies, and liability reforms that reduce litigation exposure all become viable because they address constituent pressure without validating comprehensive climate policy. Each legislative response is bounded, targeted, and framed as addressing discrete problems rather than systemic climate risk.

Legislative restraint is not dysfunction but strategic behavior under full accountability. Unified control makes congressional Republicans institutionally responsible for every outcome, which incentivizes caution. A poorly designed climate response could alienate the party’s base, invite legal challenges, or create precedents that constrain future Republican governance. Inaction carries risk, but so does action. The resulting pattern is procedural engagement without comprehensive policy commitment.

Through hearings, investigations, and appropriations constraint, congressional activity reshapes the legal and regulatory environment without extensive statutory change. The effect is not new law but altered permission structures that determine which forms of capital deployment and disclosure remain institutionally viable.

Legal and Regulatory Signal Effects

Political behavior under unified Republican control generates signals that propagate through legal and regulatory institutions, altering expectations and risk calculations even absent formal policy changes. Legal uncertainty intensifies rather than resolves. The anti-ESG campaign established ambiguity around fiduciary duty, materiality standards, and regulatory authority. The shock creates pressure for clarification, but unified Republican control signals that clarification will not favor expansive climate disclosure or ESG integration. Enforcement agencies anticipate reduced support for aggressive action. Litigants adjust strategies based on expected judicial and administrative receptivity. The shock’s visibility does not override political signals about enforcement priorities.

Regulatory agencies operate within constrained mandates and political environments. Under unified control, agencies recognize that aggressive positioning invites congressional oversight, budget constraints, and political exposure. Rulemaking processes slow. Enforcement priorities shift toward areas with clear statutory authority and minimal political controversy. Guidance documents emphasizing climate risk disclosure or ESG integration face heightened scrutiny. Where legal authority is ambiguous, agencies err toward caution. Where interpretation allows flexibility, agencies avoid positions that invite congressional investigation.

This does not mean regulatory paralysis. Agencies continue operating, rules continue being enforced, and new guidance continues emerging. But the margin of discretion narrows. The shock creates urgency for action, but political signals constrain which actions are institutionally viable. Agencies must balance constituent demand for response against political signals favoring restraint—a balance that typically resolves toward defensive positioning.

Compliance behavior responds to these signals. Firms evaluating disclosure obligations, fiduciary standards, and climate risk integration assess not only legal text but enforcement likelihood. Under unified control, firms expect reduced scrutiny for minimal compliance and increased protection from climate-related litigation. This expectation shapes how firms interpret materiality, structure disclosures, and allocate resources to climate risk management.

Judicial appointments reinforce these dynamics. Federal bench vacancies are filled by judges selected for interpretive philosophies that favor regulatory restraint and textualist statutory construction. These appointments do not immediately affect pending climate litigation, but they shape expectations about future case law development. Litigants adjust strategies. Regulatory agencies anticipate appellate outcomes. The judiciary becomes another venue where political signals translate into institutional behavior.

The aggregate effect is systematic though not uniform. Political control signals institutional tolerance for reduced climate emphasis. Different agencies, different courts, and different enforcement contexts produce variation, but the directional pressure is consistent. These signals then determine how capital allocators and corporate executives navigate the contradictory pressures of physical shock visibility and political constraint.

Capital and Firm Response under This Political Configuration

Capital allocators and corporate executives face contradictory pressures. The shock increases physical risk salience—insurance costs rise, asset values decline, and supply chain disruptions accumulate. Simultaneously, unified control signals political tolerance for reduced climate disclosure and ESG restraint. Capital cannot ignore physical risk, but political configuration determines how that risk is managed and discussed.

For capital markets, the defining condition under unified Republican control is asymmetry between physical risk visibility and regulatory signal clarity. Climate-driven losses are observable and recurring, but the legal and supervisory signals governing how those risks may be priced, disclosed, or acted upon remain constrained by political tolerance rather than financial logic.

Asset managers resolve this tension through strategic reframing. Portfolio adjustments occur, but attribution shifts. Divesting from high-emission assets becomes “energy transition positioning” rather than climate activism. Engaging companies on climate risk becomes “resilience planning” rather than ESG integration. The framing adjustment is not cosmetic but strategically necessary. Under unified control, explicit climate-based investment restrictions could invite congressional scrutiny, state-level anti-ESG legislation, or fiduciary litigation. Physical risk management continues but is embedded within broader operational, financial, and strategic frameworks that avoid ESG terminology and climate advocacy associations.

This reframing does not eliminate capital reallocation. It alters its visibility and coordination. Capital continues to migrate away from exposed assets and toward resilience-oriented investments, but it does so through individualized, opaque decision processes rather than through standardized market signals. As a result, risk is managed privately rather than priced collectively.

Pension funds face intensified versions of these pressures. Public pension boards in Republican-governed states experience direct political pressure to avoid ESG frameworks. But fiduciary duty requires addressing material risks, including those stemming from physical climate impacts. Pension funds respond by emphasizing traditional financial analysis, avoiding climate-specific mandates, and framing any climate-related adjustments as prudent risk management rather than values-based investing.

Public pension behavior illustrates the constraint clearly. Funds face climate-driven asset risk that cannot be ignored without violating prudence, yet political exposure constrains how that risk can be articulated or operationalized. The result is portfolio adjustment without narrative alignment—risk management conducted silently, without the signaling mechanisms that normally support collective market learning.

Corporate executives navigate parallel tensions through operational adaptation rather than strategic commitment. The shock creates tangible challenges—damaged infrastructure, disrupted supply chains, increased insurance costs, and regulatory uncertainty. These challenges require management attention and capital allocation. But unified control signals that comprehensive climate disclosure, ambitious emissions targets, and prominent ESG commitments may attract political scrutiny rather than institutional support.

Firms adjust accordingly. Materiality assessments narrow. Climate risk is acknowledged as financially unavoidable but not elevated to strategic priority. Sustainability reports continue but emphasize operational efficiency, resilience, and compliance rather than advocacy or voluntary commitments. Governance structures related to ESG oversight are quietly deprioritized or reframed around traditional risk management functions. This is not climate risk denial—firms experiencing direct physical impacts cannot pretend those impacts do not exist—but risk management channeled through operational adaptation, insurance, geographic diversification, and infrastructure investment rather than through comprehensive emissions reduction programs or climate advocacy.

Industry associations amplify these dynamics. Under unified control, industry organizations advocate for adaptation support, infrastructure spending, and regulatory clarity while opposing mandatory climate disclosure and emissions regulation. The shock strengthens the case for government support, but that support is framed around resilience and competitiveness rather than climate mitigation. Industry voices align with political signals, reinforcing attribution frameworks that acknowledge harm while deflecting comprehensive climate policy.

The result is not coordinated strategy but convergent behavior. Individual firms, asset managers, and industry groups independently assess political signals, legal constraints, and physical risks. Each arrives at similar conclusions—acknowledge physical impacts, manage financial exposure, avoid climate advocacy framing, and position responses within operational and resilience frameworks. The shock drives activity. Political configuration determines the institutional channels through which that activity flows.

The cumulative effect is a capital market that responds to climate risk without integrating it. Losses are absorbed, assets are repositioned, and operations are adapted, but the absence of durable legal and disclosure frameworks prevents convergence on shared valuation methodologies. Capital adjusts, but it does not coordinate.

What This Scenario Produces —and What It Does Not

This scenario produces a distinct pattern of institutional behavior under unified Republican control following a visible climate shock. Political authority is consolidated, but institutional constraint remains binding. The result is motion without settlement.

This scenario produces:

Adaptation without mitigation commitment.
Legislative and regulatory activity concentrates on infrastructure resilience, disaster preparedness, and targeted relief rather than emissions reduction. Adaptation is politically viable because it addresses visible harm while remaining compatible with the coalition’s opposition to climate mitigation mandates.

Attribution management rather than causal convergence.
The shock’s physical reality is acknowledged, but causation remains contested. Competing explanatory frames—emphasizing infrastructure failure, regulatory barriers, or international responsibility—coexist and persist. Attribution functions as a tool for managing responsibility rather than resolving disagreement.

Defensive capital and firm behavior.
Capital allocators and corporate executives continue to manage physical risk but separate that activity from ESG branding or climate advocacy. Portfolio adjustments, operational adaptation, and risk mitigation proceed, but they are framed to minimize political exposure and legal liability.

Regulatory caution under full accountability.
Agencies operate defensively. Enforcement becomes selective, guidance conservative, and rulemaking incremental. Political control does not accelerate climate-related regulation; it constrains discretion by concentrating accountability and increasing exposure to oversight and litigation.

Persistent legal uncertainty.
Unified control does not clarify fiduciary duty, materiality standards, or regulatory authority. Ambiguity established during the anti-ESG campaign remains unresolved, with litigation continuing to shape expectations unevenly across jurisdictions.

Procedural activity without comprehensive policy resolution.
Congressional hearings, investigations, and bill introductions demonstrate responsiveness but stop short of coordinated climate legislation. Procedural engagement substitutes for systemic policy commitment.

This scenario does not produce:

Resolution of underlying disagreement.
Visibility does not generate consensus. Disagreement over causation, responsibility, and institutional response persists because it is structural rather than informational.

Coordinated climate policy.
Unified control enables agenda setting but does not dissolve coalition constraints, ideological commitments, or path dependency created by prior opposition to ESG and climate disclosure.

Elimination of ESG frameworks.
Despite political pressure, ESG infrastructure endures. It remains embedded in institutional practice, contractual relationships, and international market expectations.

Market-based solutions at climate-relevant scale.
Although market mechanisms are rhetorically favored, legal uncertainty, political risk, and coalition management limit deployment of carbon pricing, large-scale renewable incentives, or economy-wide transition tools.

This scenario describes how the system processes a visible climate shock under consolidated political authority. The shock generates pressure and activity, but political configuration channels that pressure into attribution management, adaptive response, and procedural containment rather than comprehensive climate policy or institutional redesign.

This scenario describes how the system processes a visible climate shock under unified Republican control. The shock generates sustained economic and operational pressure—through insurance contraction, infrastructure strain, and localized disruption—but political authority is oriented toward limiting the scope, legitimacy, and durability of climate-risk integration rather than expanding it. Unified control enables coordinated political response, but that response is directed toward containment rather than accommodation of climate risk within financial and regulatory frameworks.

Under this configuration, legislative and executive authority are used to reinforce legal uncertainty, elevate fiduciary contestation, and expand investigatory and enforcement pressure on institutions perceived to be incorporating climate considerations beyond narrowly defined financial materiality. Regulatory action does not disappear, but it is constrained, redirected, and framed defensively. Agencies limit interpretive ambition, firms hedge exposure through procedural retreat, and capital reallocates attention toward legal and political defensibility rather than toward systematic climate-risk integration.

The result is not stasis, but a distinctive form of motion. Political action concentrates pressure upstream—through statutory redefinition, oversight, and litigation preparation—while discouraging downstream convergence in capital allocation, governance practice, or risk methodology. Institutions remain active, adaptive, and responsive to immediate constraints, yet avoid commitments that would require durable alignment around climate risk as a governing premise.

Under unified Republican control, political engagement reallocates pressure across institutions without dissolving the constraints that govern legal interpretation, capital allocation, and firm behavior. Visibility produces response, but the response channels climate stress into containment strategies that preserve discretion and contestation rather than into settled frameworks capable of coordinating long-horizon risk management.

This chapter constructs a scenario, not a forecast. It examines a plausible system response to climate change under specific, explicitly stated conditions. It does not predict outcomes, assign probabilities, or claim inevitability. Its purpose is analytical—to describe how visible climate change can coexist with institutional under-response when visibility fails to align across decision-making layers.

The scenario assumes the institutional configuration described in Chapter 6. Institutions are already operating under conditions of institutional hardening, characterized by procedural defensiveness, constrained discretion, and path-dependent risk management. Decision-making is shaped by bounded rationality. Actors optimize locally, under partial information, within mandates that privilege defensibility over reinterpretation. No sudden redesign of governance structures, information systems, or incentive frameworks is assumed.

The chapter does not assume coordinated resistance to climate action, rejection of climate science, or deliberate suppression of information. Physical climate change is treated as real, material, and increasingly observable. Weather events are visible. Climate trends are statistically visible. The binding constraint explored here is not belief, but the failure of visibility to synchronize across institutional layers in a way that compels system-level adjustment. This scenario therefore differs from accounts that attribute under-response to denial, ideology, or obstruction. Institutions respond actively to what they can see, measure, and justify. The question examined is why those responses remain localized and incremental even as physical climate risk accumulates.

Finally, this chapter is deliberately conditional. It represents one plausible system response given a particular pattern of climate impacts, timing, and salience. Different climate trajectories—even within the same seven-edge system architecture—could produce different visibility dynamics and materially alter the institutional and political scenarios examined in later chapters. This chapter establishes a baseline for that analysis, not a claim about what must occur.

Entry of Climate Shocks into the System

Climate shocks enter the system through direct physical impacts on assets, operations, and human activity. These impacts are increasingly frequent, increasingly costly, and increasingly visible at the point of occurrence. Extreme weather events—heat waves, floods, storms, wildfires—disrupt production, damage infrastructure, interrupt logistics, strain labor capacity, and generate insured and uninsured losses. Alongside these acute events, slower-moving climate trends (e.g., rising average temperatures, shifting precipitation patterns, and sea-level rise) alter baseline operating conditions over longer horizons.

Recent experience illustrates the scale of this shift. The US now experiences dozens of weather and climate disasters each year with losses exceeding $1 billion per event, a pattern that has become persistent rather than exceptional. At the same time, global climate indicators show that background conditions—average temperatures, ocean heat content, and sea levels—are moving steadily outside historical ranges, reshaping the operating environment even in the absence of headline disasters.

At the point of impact, visibility is high. Firms, households, insurers, utilities, and local governments experience climate-related disruptions directly and concretely. Facilities shut down. Supply chains reroute. Insurance claims are filed. Maintenance costs rise. Emergency expenditures increase. These events are not abstract signals; they are operational facts.

In high-exposure regions, this visibility has translated into immediate operational strain. Insurers have faced surges in claims tied to wildfire, flood, and storm losses, while utilities and municipalities have confronted repeated repair and recovery cycles within compressed timeframes. These impacts are experienced first as operational stress rather than as strategic or systemic signals.

However, these impacts enter the system unevenly and asynchronously. Climate shocks are geographically concentrated, sector-specific, and temporally irregular. Some regions experience repeated disruptions while others remain relatively insulated. Some industries face acute exposure, while others encounter marginal or delayed effects. Extreme events cluster episodically rather than occurring in synchronized waves. This uneven distribution prevents immediate system-wide recognition of a shared condition.

The result is a pattern in which localized escalation coexists with apparent aggregate stability. Even as certain states, sectors, or communities confront repeated disruption, national or sector-wide indicators may continue to appear manageable, reinforcing the perception that impacts remain contained rather than systemic.

A critical distinction shapes early system response. Weather events are visible, climate trends are statistically visible, but meaning is not yet shared. Individual events are experienced as discrete disruptions. Long-term trends are observable through data and models, but they do not yet register as common operational premises across institutions. The system encounters climate change first as variability rather than as direction.

For example, rising average temperatures, accelerating sea-level rise, and intensifying precipitation extremes are well-documented in scientific assessments, yet these trends often remain framed as background conditions rather than as drivers requiring coordinated institutional reinterpretation. Data exist, but they have not yet reorganized decision baselines across governance, finance, and operations.

At this stage, climate shocks are typically managed as operational contingencies. Units closest to disruption respond using tools designed for continuity and recovery rather than diagnosis or reinterpretation. Schedules are adjusted, inventories rebalanced, contracts renegotiated, repairs undertaken. These responses are rational, immediate, and effective within their scope. Escalation beyond the operational layer is limited because existing frameworks treat disruption as episodic rather than structural.

As impacts move inward, they are translated into domain-specific metrics that allow institutional processing. Physical damage becomes insurance losses, capital expenditures, or maintenance costs. Disruptions appear in financial statements as volatility, write-downs, or margin pressure. Public-sector impacts surface as emergency spending or infrastructure repair budgets. Each translation is necessary for institutional action, but each also strips away contextual information about causality, recurrence, and trajectory.

In the insurance and public finance domains, this translation has been particularly visible. Climate-driven losses increasingly appear as pricing adjustments, coverage withdrawals, or budgetary stress rather than as explicit indicators of rising physical risk. The system registers financial consequences more readily than it integrates physical causation.

Initial framing matters. When climate-related disruptions are categorized as isolated events, regional anomalies, or routine operational risks, they do not trigger cross-institutional synthesis. The system absorbs shocks incrementally, resolving each instance within existing practices without requiring a shared interpretation of cause or future significance. Climate change is visible at entry, but it enters the system fragmented—experienced intensely in some places, abstractly in others, and not yet as a common condition demanding coordinated response.

Fragmentation and Translation of Climate Signals

As climate shocks move beyond the point of impact, their informational content is reshaped by the institutional pathways through which they travel. Physical disruption does not propagate as a single signal. It is translated into domain-specific representations that are legible within existing organizational, financial, and regulatory frameworks. This translation enables response, but it also fragments meaning.

Operational units record climate impacts as downtime, repair schedules, logistics rerouting, or productivity losses. Finance functions register the same events as volatility, claims expenses, capital expenditures, or earnings variance. Risk teams convert them into adjustments to pricing assumptions, exposure limits, or loss models. Public entities reflect impacts through emergency appropriations, infrastructure repair budgets, or service disruptions. Each representation is internally coherent. None is false. But each captures only a slice of the underlying physical reality.

Fragmentation arises because translation prioritizes compatibility over completeness. Information is reshaped to fit existing categories, thresholds, and reporting conventions. Context that cannot be expressed in those terms—such as recurrence risk, nonlinear escalation, or interaction with longer-term climate trends—is attenuated or lost. The system does not suppress information; it reformats it.

This process is reinforced by institutional specialization. Different domains observe climate impacts through distinct functional lenses, each optimized for local decision-making. Operations focus on continuity. Finance emphasizes quantifiable exposure. Governance attends to compliance and oversight. Insurance centers on loss attribution and pricing. These perspectives do not naturally converge. They move in parallel, connected loosely by accounting and reporting interfaces rather than by shared interpretive frameworks.

As signals travel outward, visibility declines even as data volume increases. Local impacts are vivid and detailed; translated metrics are abstract and standardized. Repetition across locations does not automatically generate synthesis because each instance is processed independently within its domain. What might appear, in aggregate, as an emerging pattern remains distributed across institutional silos.

Fragmentation is further amplified by timing. Data are collected episodically, reviewed on fixed cycles, and reported with delay. Operational disruptions are addressed immediately; financial implications appear later; and governance responses follow established reporting and review schedules. By the time information reaches higher-level forums, it has already been normalized, averaged, or contextualized as routine variance.

The result is a system that knows many things without knowing the same thing. Climate impacts are observed, recorded, and acted upon, but not integrated into a shared understanding of direction or scale. Fragmentation does not reflect ignorance or denial. It reflects the way complex systems translate physical reality into actionable inputs under bounded mandates.

This translation process sets the conditions for what follows. Once climate signals are fragmented, attribution becomes contested, aggregation dilutes directionality, and institutional response defaults to incremental adjustment rather than reinterpretation. The system remains responsive, but only within the limits imposed by how information is rendered legible.

Attribution Limits under Bounded Rationality

Fragmented visibility constrains not only synthesis but attribution. Even when climate-related disruptions are clearly observed, institutions face persistent difficulty determining why they occurred, how much weight to assign to physical drivers, and whether those drivers warrant a change in underlying assumptions. Attribution failure in this scenario is not denial of physical events. It is uncertainty about causal priority.

Weather events are visible. Extreme heat, flooding, wildfire, drought, and storm damage are experienced directly and documented extensively. Climate trends are also statistically visible, observable through temperature records, sea-level rise, shifting precipitation patterns, and increasing frequency or severity of certain events. What remains contested is not whether these phenomena exist, but how decisively they should be treated as causal inputs in institutional decision-making.

Physical impacts rarely present as single-cause events. A disrupted supply chain may plausibly be attributed to weather, infrastructure condition, labor availability, or management decisions. Insurance losses may reflect climate exposure, but also pricing assumptions, legacy portfolios, or regulatory constraints. Infrastructure failure may involve physical stress interacting with deferred maintenance or design standards. Climate change operates as a background amplifier, complicating attribution rather than replacing proximate explanations.

Under bounded rationality, institutions act on causes that are legible, actionable, and defensible within their mandates. Proximate explanations dominate because they align with existing tools and responsibilities. Distal or systemic causes—especially those unfolding over long horizons—are harder to isolate, harder to justify, and harder to operationalize. Institutions therefore hesitate to treat climate drivers as decisive unless attribution crosses a high evidentiary threshold.

This hesitation is reinforced by asymmetry of risk. Treating climate change as causally decisive can require revising assumptions, reallocating capital, or altering governance frameworks in ways that are difficult to defend if attribution remains contested. In contrast, responding to proximate causes allows institutions to act decisively without reopening foundational interpretations. Bounded rationality thus favors explanations that minimize interpretive exposure rather than those that maximize causal completeness.

Attribution is further constrained by temporal mismatch. Establishing climate causation often requires longitudinal analysis that unfolds more slowly than operational or financial decision cycles. By the time causal clarity improves, decisions have already been made, disruptions have been absorbed, and attention has shifted. Learning arrives late, after the window for reinterpretation has closed.

As a result, climate change is acknowledged but rarely centered. It appears as one factor among many, influencing outcomes without commanding interpretive priority. Institutions manage its effects without fully integrating its implications. Under-response emerges not because climate change is invisible, but because its causal role remains insufficiently decisive within the limits of bounded rationality.

This attribution gap is a critical hinge in the scenario. It explains why visible physical disruption does not automatically translate into systemic adjustment, and why institutions continue to operate within established frameworks even as climate-related stress accumulates. What follows depends not on disbelief, but on how attribution interacts with aggregation and institutional response under partial visibility.

Aggregation Lag and Macro Dilution

As climate-related disruptions propagate beyond the operational layer, they are aggregated into system-level indicators that emphasize stability, comparability, and continuity. This process introduces delay and dilution. Localized disruptions are standardized, averaged, and offset across sectors and regions, producing macro signals that lag behind physical reality.

Aggregation is not neutral. It is designed to support governance, risk management, and comparability across heterogeneous contexts. In doing so, it smooths variance and suppresses outliers. Losses in one region are offset by gains elsewhere. Sectoral stress is absorbed into broader performance measures. What emerges is a macro picture that remains legible and orderly even as underlying conditions deteriorate unevenly.

Recent climate data illustrate this smoothing effect clearly. The US has experienced a sustained increase in billion-dollar weather and climate disasters, with dozens of such events occurring annually and cumulative losses rising sharply over the past decade. Yet these losses are distributed across regions, sectors, and balance sheets in ways that prevent any single year or location from dominating national economic indicators. The result is a pattern in which physical disruption intensifies while macro aggregates continue to signal overall economic resilience, reinforcing the appearance of continuity even as exposure grows.

This lag is especially pronounced for climate-related risks because impacts are geographically concentrated and sector-specific. Repeated disruptions in high-exposure regions—coastal flooding, wildfire-prone areas, heat-stressed infrastructure corridors—do not immediately alter national averages or headline indicators. The system registers stress locally while preserving apparent stability at scale.

Aggregation lag differs from attribution failure. Attribution concerns uncertainty about causation; aggregation concerns signal transmission. Even when institutions understand climate drivers locally, the time required for those observations to influence macro indicators delays institutional recalibration. Climate risk enters the macro layer slowly, filtered through accounting cycles, regulatory thresholds, and reporting conventions that privilege historical baselines over forward-looking exposure.

This dilution is visible in climate-sensitive sectors such as insurance and housing. While insurers and reinsurers increasingly adjust pricing, coverage, and underwriting in high-risk regions, these shifts initially appear as localized market frictions rather than as macroeconomic signals. Coverage withdrawals, premium increases, and non-renewals accumulate unevenly, affecting specific communities long before they register in national housing or financial stability metrics.

Federal assessments of the US insurance and housing markets confirm this pattern. Climate-driven non-renewals, premium increases, and coverage gaps are concentrated in high-risk regions, particularly in coastal and wildfire-exposed states. At the same time, outdated flood mapping and slow revision cycles mean that much of this rising exposure remains invisible to national risk frameworks. Development, lending, and infrastructure investment continue under assumptions derived from historical conditions, allowing physical risk to compound beneath macro indicators that still signal relative stability.

Similarly, public-sector aggregation mechanisms lag physical change. Flood exposure continues to be assessed using historical data frameworks even as forward-looking climate models indicate rising risk. Outdated flood maps and slow revision cycles allow development, lending, and infrastructure investment to proceed under assumptions that no longer reflect emerging physical conditions. The macro system preserves continuity by design, even as exposure grows beneath the surface.

The result is a persistent gap between physical reality and macro-level recognition. Climate-related losses are real, recurring, and increasingly costly, yet their system-wide implications remain muted by aggregation practices that prioritize stability and comparability over early signal detection. Macro indicators do not deny climate risk. They delay its visibility.

This delay reinforces the conditions described in Chapter 6. Institutions relying on macro signals encounter insufficient justification for reinterpretation. Governance baselines remain anchored to historical performance. Political salience remains muted until stress becomes widespread and synchronized. Aggregation preserves stability—until it no longer can.

Institutional Response under Partial Visibility

Under conditions of fragmented signals, attribution limits, and aggregation lag, institutional response is shaped less by the magnitude of physical disruption than by what can be observed, justified, and defended within existing frameworks. Institutions do not fail to respond. They respond in ways that are consistent with partial visibility and constraint-rational behavior.

Institutions act on what is legible within their mandates. Risks that are clearly attributable, widely recognized, and compatible with established metrics receive attention and resources. Risks that are diffuse, probabilistic, or difficult to isolate are managed indirectly or deferred. This produces a pattern of incremental adjustment rather than systemic reinterpretation.

Responses therefore take familiar forms. Procedures are refined. Thresholds are tightened. Documentation requirements expand. Risk classifications are updated at the margin. These actions address observed effects without reopening underlying assumptions about exposure, causation, or long-term trajectory. Activity continues, but it is channeled through increasingly standardized pathways.

Recent institutional behavior in climate-exposed sectors illustrates this pattern. Insurance markets, for example, have responded to rising catastrophe losses through pricing adjustments, coverage restrictions, non-renewals, and selective market withdrawal in high-risk regions. Regulators and supervisory bodies have increased data collection, reporting requirements, and stress-testing exercises. These responses acknowledge physical risk and manage near-term exposure, but they stop short of forcing system-wide reinterpretation of long-term insurability, land use, or capital allocation. The emphasis remains on managing observable impacts rather than redesigning underlying assumptions.

This behavior is boundedly rational. Institutions optimize locally under incomplete information, asymmetric downside risk, and accountability structures that reward defensibility over foresight. Escalation beyond visible evidence appears unjustified. Acting too aggressively on partial signals carries reputational, legal, and political risk that exceeds perceived benefit.

Importantly, under-response does not reflect denial of climate change or rejection of physical reality. Climate impacts are acknowledged, tracked, and managed where they are directly experienced. What is absent is coordinated reinterpretation across decision-making layers. Without shared visibility, institutional responses remain compartmentalized.

Institutional hardening amplifies this pattern. As discretion has already been displaced into procedures, the system defaults to formal reinforcement rather than interpretive expansion. New information is processed through existing rules instead of prompting their reconsideration. Responses stabilize behavior without revisiting premises.

This procedural orientation is reinforced by oversight and accountability structures. Institutions are evaluated on whether established processes were followed, disclosures were made, and controls were applied—not on whether emerging risks were anticipated early or reframed strategically. As a result, climate-related actions that can be documented, audited, and defended are favored over those that require interpretive judgment under uncertainty. The system signals responsibility and diligence even as cumulative physical risk continues to outpace institutional adaptation.

This produces a distinctive response profile. Institutions appear active and responsible. They demonstrate compliance, diligence, and responsiveness to documented risks. Yet their actions do not scale with cumulative physical change. Adjustment remains reactive rather than anticipatory, localized rather than systemic.

Partial visibility also constrains coordination. Escalation requires shared evidence that crosses domains. In its absence, no actor has sufficient authority or justification to initiate system-level change. Each institution acts appropriately within its remit, while the system as a whole under-responds.

The result is not paralysis, ignorance, or obstruction. It is a stable mode of functioning under uncertainty. Institutions continue to manage risk, but they do so in ways that preserve existing structures and defer reinterpretation. Climate stress is absorbed through incremental adaptation rather than transformative response. This pattern persists until visibility changes—through shock synchronization, attribution clarity, or aggregation breakdown. Until then, institutional response remains disciplined, constrained, and insufficient relative to cumulative physical risk.

Temporal Accumulation without Narrative Convergence

Over time, physical climate impacts accumulate. Assets are stressed repeatedly, recovery windows shorten, insurance losses compound, and adaptation costs rise. Yet accumulation alone does not produce convergence in interpretation. Repetition does not guarantee synthesis.

In this scenario, each disruption is managed as a discrete event rather than as evidence of a shared trajectory. Floods, heatwaves, fires, storms, and infrastructure failures are experienced as episodic shocks, even when their frequency increases. Institutions respond to each instance using established tools, resolving immediate effects without revisiting underlying exposure or long-term risk.

As disruptions recur, their informational content declines. What initially appears exceptional becomes expected. Variance is absorbed into baselines. Adjustments made after earlier events normalize future ones. Adaptation, rather than clarifying risk, can obscure it by embedding response into routine operations.

This pattern is reinforced by well-documented cognitive and institutional dynamics in which repeated exposure to adverse conditions leads to normalization rather than escalation. As disruptions become familiar, their perceived informational value declines, even as their objective frequency and severity increase. Repetition reduces salience, encourages incremental adaptation, and weakens the likelihood that cumulative change will be interpreted as requiring structural response rather than continued adjustment.

This normalization process weakens narrative formation. Without synchronization across regions, sectors, or time horizons, no single event compels reinterpretation. Losses are uneven. Impacts are staggered. Benefits of adaptation are localized. The system experiences stress without shared meaning.

Temporal accumulation also interacts with institutional incentives. Because each event is addressed successfully at the local level, escalation appears unnecessary. Effective short-term response reinforces confidence in existing frameworks, even as cumulative exposure grows. Institutions learn how to cope, not how to reframe.

Narrative convergence requires more than repetition. It requires alignment across visibility, attribution, and aggregation. In this scenario, those conditions do not materialize. Climate change remains visible as weather and statistically observable as trend, but not legible as a unified system-level signal demanding reinterpretation. The result is a paradoxical condition. Physical risk intensifies while interpretive consensus stalls. Institutions adapt repeatedly without coordinating meaning. Time passes, impacts accumulate, but the system’s understanding does not fundamentally change.

This condition is stable precisely because it is functional. The system continues to operate. Losses are managed. Recovery occurs. No single failure forces a reckoning. Yet the absence of narrative convergence leaves the system vulnerable to future shocks that exceed its adaptive capacity.

Temporal accumulation without convergence sets the stage for what follows. It explains why later political and institutional responses unfold within hardened structures rather than transforming them. It clarifies why escalation, when it occurs, is abrupt rather than gradual.

What accumulates here is not shared understanding, but latent stress. The system moves forward, but without a common story about where it is going—or why.

Why This Is a Scenario, not a Prediction

This chapter constructs a scenario, not a forecast. It does not assert that this pattern of institutional response will occur, nor that it is the most likely outcome. It describes one coherent way a system characterized by institutional hardening and bounded rationality can respond to climate change when physical impacts are visible, but system-level meaning remains misaligned.

The scenario depends on specific conditions holding simultaneously: uneven climate impacts, contested attribution, delayed aggregation, and institutions operating within hardened procedural constraints. Under those conditions, under-response is not irrational or ideological. It is constraint-rational.

Different climate trajectories could produce different outcomes. More synchronized shocks, sharper regional concentration, clearer attribution, or faster aggregation could accelerate visibility and force reinterpretation. Alternative data salience—through insurance markets, infrastructure failure, fiscal exposure, or public health impacts—could destabilize this configuration sooner or more decisively.

Even within the same seven-edge system architecture, variation in timing, magnitude, and distribution of climate impacts could materially alter institutional behavior. A different sequence of shocks could compress learning cycles, overcome procedural defensiveness, or force coordination across decision layers that remains absent in this scenario. For this reason, the scenario presented here should be read as conditional rather than deterministic. It identifies a plausible mode of system behavior under a specific configuration of constraints. It does not claim inevitability.

This matters for what follows. The political and institutional scenarios examined in Chapters 8–11 are constructed on the assumption that institutional hardening remains in place and that climate impacts continue to accumulate without producing immediate narrative convergence. Those chapters explore how stress is reallocated within a hardened system, not how the system is redesigned.

If climate visibility intensifies faster than assumed here, or if attribution becomes politically unavoidable, the boundary conditions of those scenarios would shift. The analysis would still apply, but the trajectories would differ. The purpose of this chapter is therefore not to predict outcomes, but to discipline imagination. It clarifies what kinds of responses are plausible given the constraints already in place and what kinds of change would require those constraints to break.

Part III: Baseline Conditions and Climate Stress

Part III establishes the baseline against which all subsequent scenarios are evaluated. It does not yet introduce political change or alternative governance pathways. Instead, it describes how the anti-ESG ecosystem is currently configured and how that configuration responds when climate stress enters the system without forcing immediate reinterpretation.

Chapter 6 is not a forecast. It is a present-state description. It characterizes the existing configuration of institutional behavior under persistent uncertainty, using the analytical architecture and constants established in Part II. The chapter shows how institutions across domains converge on similar patterns of action—not through coordination or ideological agreement, but through repeated, constraint-rational adaptation. Institutional hardening describes a system that remains active, orderly, and legible while progressively narrowing discretionary space. It explains why stability can emerge without resolution, why behavior can converge without consensus, and why procedural defensibility becomes the dominant mode of institutional response.

Chapter 7 then introduces climate stress into that already-hardened system. Unlike later political scenarios, it does not posit a change in governance, electoral outcomes, or institutional authority. It constructs a conditional scenario grounded in observed and projected climate impacts, asking how a system characterized by institutional hardening responds when physical climate change is real, material, and increasingly visible but not yet synchronized across decision-making layers.

The key analytical move in Chapter 7 is not to question whether climate change is occurring, but to examine why visibility alone does not guarantee system-level adjustment. Climate shocks enter the system through operational disruption and financial loss, but they are translated, fragmented, and aggregated in ways that delay shared interpretation. Institutions respond actively and rationally within their mandates, yet those responses remain localized and incremental. The result is under-response without denial, adaptation without reinterpretation, and activity without convergence.

Together, Chapters 6 and 7 establish the system’s baseline condition under stress. They show how a hardened institutional configuration absorbs pressure, manages disruption, and preserves functionality while deferring deeper recalibration. This baseline is essential. It defines what kinds of change are possible quickly, what kinds require constraint failure, and why later political scenarios do not begin from a neutral or flexible starting point.

Part IV builds directly on this foundation. The political scenarios that follow do not reset the system. They explore how stress is reallocated when electoral outcomes alter authority and incentives within a system that has already learned how to protect itself from uncertainty—at the cost of flexibility, interpretive ambition, and discretionary reach.

Institutional Hardening

Institutional hardening describes a present-state system configuration, not a future forecast or a normative judgment. It characterizes how institutions are currently behaving under persistent uncertainty within the system architecture defined in earlier chapters. The scenario does not assume coordination, ideological convergence, or strategic intent. It describes an emergent pattern produced by repeated, rational adaptation to unresolved constraint.

This scenario is not a prediction about what institutions will do. It is an analytical description of what institutions are already doing, conditional on the constants established in Chapter 5. The analysis that follows assumes no sudden legal resolution, no collapse of political contestation, and no decisive clarification of fiduciary or reputational standards. Under those conditions, institutions converge toward a shared behavioral posture without ever agreeing on its causes or consequences.

Institutional hardening is therefore not a condition of withdrawal or paralysis. Institutions remain active across all seven edges of the system. Capital continues to move. Governance continues to function. Operations continue. Political activity persists. What changes is the mode of institutional action. Discretion is progressively constrained. Decision-making is reorganized around defensibility rather than interpretive ambition. Stability is achieved through procedural narrowing rather than consensus.

This chapter does not argue that institutional hardening is optimal, inevitable, or permanent. It establishes it as a baseline configuration—the system state against which subsequent scenarios are evaluated. If the reader rejects this configuration, the logic of later chapters will not follow. That disagreement is legitimate, but it must be explicit.

How Institutional Hardening Emerges from the Seven-Edge System

Institutional hardening does not originate at any single point in the system. It emerges from the cumulative effect of pressure transmitted across all seven structured transmission pathways, each operating according to its own logic and constraints. No pathway dominates the others, and none carries inherent causal priority. Hardening arises because each pathway independently rewards defensibility, pre-commitment, and proceduralization under conditions of persistent uncertainty.

Along the Narrative → Law pathway, plural and unresolved narratives are translated into narrow forms of legal legibility. Competing accounts of responsibility, risk, and legitimacy do not prevent action, but they raise the reputational and legal cost of interpretive choice. Legal systems extract only those elements of narrative that can be abstracted, stabilized, and enforced, privileging formulations that minimize ambiguity rather than resolve disagreement.

Across the Law → Capital Allocation pathway, legal uncertainty is converted into permission risk. Capital does not withdraw wholesale, but its deployability narrows. Anticipated enforcement, jurisdictional variation, and ambiguous fiduciary standards absorb risk capacity, encouraging allocation decisions that can be defended ex ante rather than optimized ex post. Capital continues to move, but within increasingly constrained and standardized parameters.

Through the Capital Allocation → Firm Governance pathway, external constraint is internalized as procedural architecture. Boards, committees, approval thresholds, and documentation requirements expand as mechanisms for containing exposure. Governance becomes the primary site where uncertainty is managed, not by resolving competing claims, but by embedding caution into decision rules.

As pressure moves from Firm Governance → the Real Economy, discretionary judgment is displaced rather than eliminated. Judgment migrates upstream into the design of policies, thresholds, and escalation pathways, where it can be exercised prospectively and justified defensively. Once embedded in process architecture, execution is deliberately rendered judgment-poor. Operational activity continues, but through narrower channels and slower pathways shaped by compliance and review.

Across the Real Economy → Macro Conditions pathway, these micro-level frictions aggregate into indicators of stability rather than resolution. Output persists, markets function, and activity appears orderly, even as adaptability and responsiveness decline. Macro signals reflect continuity, masking the accumulation of rigidity beneath the surface.

Through the Macro Conditions → Politics pathway, apparent stability dampens urgency without eliminating contestation. Aggregate indicators do not converge on clear diagnoses, and distributional effects remain uneven. Political actors interpret persistent ambiguity and volatility as reasons to reduce discretionary exposure rather than to arbitrate underlying disagreement.

Finally, along the Politics → Narrative Entrepreneurs pathway, unresolved contestation replenishes narrative plurality. Political outcomes and attention signals authorize certain frames while foreclosing others, reshaping which narratives are produced and amplified in the next cycle. Rather than restoring coherence, political conflict feeds back into narrative competition in altered form.

Institutional hardening emerges because each transmission pathway independently selects for defensibility under uncertainty and because those selections accumulate faster than opportunities for learning, coordination, or reversal. Discretion is not removed from the system; it is relocated into upstream design choices that narrow downstream flexibility. The result is a system that continues to function, but does so on tightened ground—stable in appearance, constrained in adaptation.

Constraint Accumulation as the Dominant Mechanism

The defining mechanism of institutional hardening is constraint accumulation, not ideological alignment or strategic retreat. Institutions do not converge because they share beliefs. They converge because similar adaptations are repeatedly selected under similar conditions.

Each individual adjustment is modest : a revised guideline, an expanded disclosure requirement, a n additional review step, a narrower eligibility definition, a more conservative interpretive memo. In isolation, none of these changes appears transformative. Each is easily justified as prudent, temporary, or technical.

Over time, however, these adjustments layer. Discretion becomes costly not because it is prohibited, but because it is difficult to defend across multiple evaluative frames. The burden of justification increases even when decisions are internally coherent. Institutions respond by shifting judgment upstream—away from individual decisions and into the design of processes.

Procedural defensiveness functions as a self-reinforcing internal selection mechanism. Compliance frameworks privilege repeatability over interpretation. Oversight functions reward adherence to process rather than contextual judgment. Training emphasizes rule application rather than evaluative reasoning. Performance evaluation favors consistency and defensibility over situational optimization. Together, these mechanisms select for further procedural reinforcement, making deviation increasingly salient and costly. The system does not merely adopt defensiveness; it continually reproduces it.

Constraint accumulation operates through local feedback loops. Scrutiny generates caution. Caution generates procedural reinforcement. Reinforced procedures reduce exposure. This loop does not require coordination or communication. It is selected for because it reduces institutional vulnerability under persistent uncertainty. The result is a system that remains active and legible, but increasingly formalized. Decision-making is displaced from context-specific evaluation to rule-bound application. Institutions appear stable not because disagreement has been resolved, but because it has been procedurally contained.

This is the internal logic of institutional hardening: A configuration stabilizes behavior without resolving conflict, preserves functionality without restoring interpretive openness, and maintains institutional presence without arbitrating underlying disagreement—achieving coherence through procedural containment rather than interpretive resolution.

How Institutional Hardening Manifests Across the Seven Edges

Institutional hardening does not propagate through new mechanisms. It expresses itself through the system’s existing edges, altering how pressure is transmitted, processed, and absorbed at each transition. What changes is not the direction of influence, but the mode by which institutions respond. Across the system, discretion is progressively displaced by formalization, interpretive flexibility by procedural defensibility, and contextual judgment by standardized criteria.

E1 —Narrative → Law: From Interpretive Pluralism to Definitional Narrowing

Under institutional hardening, the narrative-to-law edge becomes increasingly selective. Legal institutions continue to receive narrative pressure, but they respond by extracting only those elements that can be rendered precise, bounded, and enforceable. Broad interpretive claims are narrowed into definitional disputes. Ambiguity is treated as liability rather than as space for judgment.

As narratives intensify and diverge, legal actors prioritize legibility over completeness. Categories are stabilized not because they resolve disagreement, but because they reduce exposure. Legal responses increasingly take the form of definitional constraints, scope limitations, and procedural triggers rather than expansive mandates or principles. What hardens is not the content of law, but its tolerance for interpretive openness.

Illustrative example

As ESG narratives became politically contested across US states, legal institutions did not attempt to reconcile competing accounts of risk, responsibility, or legitimacy. Disagreement persisted rather than converging toward a shared interpretation, increasing the reputational and legal cost of adopting expansive or principle-based legal positions.

Instead, narrative pressure was selectively translated into narrow definitional questions that could be rendered administrable within existing legal frameworks—such as whether particular considerations qualified as pecuniary, whether certain practices constituted prohibited boycotts, or whether disclosure thresholds had been crossed. These abstractions appeared repeatedly in state statutes, attorney general opinions, and fiduciary guidance. What crossed the edge was not the narrative itself, but a reduced operational proxy that allowed legal action to proceed while limiting interpretive exposure.

E2 —Law → Capital Allocation: From Permissibility to Conservative Eligibility

At the law-to-capital edge, institutional hardening manifests as anticipatory conservatism. Capital institutions do not wait for definitive legal resolution. Instead, they translate legal ambiguity into risk parameters that bias decision-making toward defensibility. Permissible actions are treated as risky if justification would be costly, contested, or jurisdictionally complex.

Eligibility screens, contractual representations, and internal compliance thresholds increasingly substitute for case-by-case judgment. Capital continues to flow, but within narrower corridors defined by standardized interpretations of legal exposure. What hardens is not capital supply, but the conditions under which capital can be deployed without escalating institutional risk.

Illustrative example

Following the passage of Texas anti-ESG statutes affecting municipal contracting, several major banks reduced or exited participation in public finance underwriting in the state. This response did not reflect a rejection of ESG considerations as an investment philosophy, but an assessment that participation under newly contested eligibility rules carried heightened legal and reputational risk.

For institutions that sought to preserve or re-establish access to state business, adjustment occurred through internal constraint rather than outward policy change. Eligibility screens were tightened, contractual representations were revised, and documentation requirements were expanded to ensure that participation could be defended ex ante under uncertain enforcement conditions. Legal ambiguity was thus translated into capital constraints: Capital continued to flow, but only where exposure could be rendered procedurally defensible. What crossed the edge was not a directive to alter investment objectives, but a redefinition of the conditions under which capital deployment remained permissible without escalating institutional risk.

E3 —Capital Allocation → Firm Governance: From Strategic Optionality to Procedural Control

As capital constraints formalize, firm governance adapts by internalizing defensibility as a priority. Governance frameworks shift emphasis from strategic discretion to risk containment. Decision authority is redistributed toward committees, controls, and review processes designed to demonstrate procedural rigor rather than contextual responsiveness.

Boards and senior management continue to make decisions, but those decisions are increasingly shaped upstream by capital-linked expectations concerning disclosure, oversight, and risk management. Governance hardens by converting external constraints into internal rules—policies, escalation paths, and approval thresholds—that narrow the space for discretionary interpretation at the point of action.

Illustrative example

As capital allocators adjusted stewardship expectations, disclosure standards, and risk signaling under heightened legal and political scrutiny, firms responded by reorganizing internal governance rather than by altering stated strategic objectives. Sustainability-related decisions were increasingly routed through expanded board committees, additional review layers, and formal escalation pathways designed to demonstrate procedural rigor.

These governance adjustments did not prescribe substantive outcomes. Instead, they restructured how decisions were evaluated, justified, and approved. By embedding defensibility into committee mandates, documentation requirements, and approval thresholds, firms converted external capital-related uncertainty into internal process constraints. What crossed the edge was not a demand to pursue or abandon particular strategies, but pressure to govern decisions in ways that could be defended to capital providers under contested conditions.

E4 —Firm Governance → Real Economy: From Policy Intent to Operational Filters

At the governance-to-real-economy edge, institutional hardening appears as friction rather than directive. Governance priorities are operationalized through internal filters—budgeting criteria, procurement standards, performance metrics—that shape which projects advance and which stall. Operational actors experience these changes as revised thresholds, extended timelines, or additional justification requirements rather than as explicit strategic shifts.

This hardening does not eliminate activity. It redirects it. Investments with clear, short-term defensibility proceed more easily than those requiring contextual explanation or longer-horizon justification. Over time, operational decision-making becomes more standardized and less exploratory, even in the absence of changes in market demand or formal regulation.

Illustrative example

In the US insurance sector, enhanced board-level climate risk governance translated into revised underwriting standards, pricing adjustments, and exposure limits. Insurers did not issue blanket prohibitions on coverage or announce strategic withdrawal from climate-exposed markets. Instead, governance priorities were operationalized through revised underwriting criteria, expanded documentation requirements, and more stringent review processes.

These operational shifts did not arise from changes in customer demand or formal regulatory mandates. They reflected the application of internal filters that increasingly favored risks that could be justified procedurally under heightened governance scrutiny. Coverage continued, but activity was redirected toward regions, asset classes, and risk profiles that were easier to defend ex ante, illustrating how governance intent shaped real-economy outcomes indirectly through operational constraint rather than directive.

E5 —Real Economy → Macro Conditions: From Distributed Caution to Aggregate Rigidity

As hardened operational practices accumulate, their effects register at the macro level through aggregation. Individual decisions—delayed investments, extended asset lifespans, conservative risk pricing—are not macroeconomic in isolation. In combination, they contribute to patterns of slower adjustment, regional divergence, and increased sensitivity to shocks.

Macro indicators reflect these effects without preserving their institutional origins. What appears as reduced flexibility or heightened volatility at the macro level is often the downstream manifestation of widespread micro-level defensiveness. Institutional hardening thus contributes to macro conditions that are stable yet brittle, resilient in routine operation but less adaptive to disruption.

Illustrative example

Climate-driven insurance nonrenewals and premium increases concentrated in high-risk regions such as Florida and California reduced coverage availability and affordability at the household and firm level. These decisions were operational responses to underwriting defensibility and capital adequacy constraints rather than coordinated attempts to alter regional economic outcomes.

In aggregate, however, these localized adjustments propagated into broader housing market stress, credit contraction, and regional economic fragility. The resulting macro condition did not originate from macroeconomic policy choices, but from the accumulation of micro-level defensive responses selected under hardened institutional conditions.

Climate-driven insurance nonrenewals and premium increases concentrated in high-risk regions such as Florida and California reduced coverage availability and affordability at the household and firm level. These actions were taken as localized operational responses to underwriting defensibility, loss volatility, and capital adequacy constraints, not as coordinated efforts to influence regional economic outcomes.

In aggregate, however, these dispersed adjustments propagated into broader housing market stress, tighter credit conditions, and regional economic fragility. What emerged at the macro level was not the result of a single policy choice or strategic decision, but the arithmetic accumulation of many micro-level defensive responses selected under conditions of institutional hardening, illustrating how real-economy caution translates into macro rigidity through aggregation rather than intent.

E6 —Macro Conditions → Politics: From Structural Stress to Salience Compression

When hardened macro conditions become visible, political response is shaped by salience rather than diagnosis. Aggregate stress—rising costs, constrained insurance availability, regional disparities—compresses political time horizons and elevates demands for immediate protection. Longer-horizon risk management becomes easier to frame as optional or expendable.

Institutional hardening amplifies this effect by limiting the range of politically legible responses. Because many underlying constraints are procedural and indirect, political actors focus on visible pressure points rather than upstream institutional design. What hardens here is not political ideology, but the incentive structure governing what kinds of responses are rewarded.

Illustrative example

Periods of elevated inflation and cost-of-living pressure altered political salience by compressing voter time horizons and prioritizing immediately felt economic stress. Rising prices, insurance availability constraints, and visible affordability shocks became dominant reference points for political attention, even when their underlying causes were diffuse and structurally complex. These conditions elevated demands for short-term relief and identifiable intervention targets, while longer-horizon risk management initiatives struggled to maintain visibility.

The transmission across this edge operated through salience rather than diagnosis. Political actors and media frames translated aggregate macro stress into simplified narratives of blame, urgency, and protection that were electorally legible. Structural explanations requiring coordination across legal, capital, or governance domains remained difficult to sustain, not because they lacked validity, but because hardened institutional constraints limited the range of responses that could be rendered immediately visible and attributable. As a result, political incentives shifted faster than institutional capacity to adapt, reinforcing salience-driven action over systemic reinterpretation.

E7 —Politics → Narrative Entrepreneurs: From Interpretive Exploration to Demand-Side Optimization

At the final edge, political outcomes feed back into narrative production under hardened conditions. Narrative entrepreneurs respond less to analytical completeness than to demonstrated resonance. Electoral signals, engagement metrics, and funding flows indicate which frames are politically viable. Over time, these signals favor simplified, high-certainty narratives that align with hardened institutional incentives.

Narrative production remains active and competitive, but it becomes increasingly optimized for transmission rather than exploration. Complex institutional dynamics are compressed into accessible causal stories. What hardens is the narrative supply itself: Fewer frames are tested, and those that succeed are refined rather than reconsidered.

Illustrative example

Following state-level anti-ESG actions and the publication of financial-institution restriction lists, narrative production shifted toward frames that had demonstrated political and electoral traction. ESG practices were increasingly characterized as forms of economic boycott, financial coercion, or ideological exclusion—not because new empirical evidence about governance or investment behavior had emerged, but because these frames had been validated through legislative success, media amplification, and constituent engagement.

The transmission across this edge operated through demand-side signals rather than analytical refinement. Polling results, primary outcomes, fundraising patterns, and audience engagement metrics indicated which narratives resonated within hardened political conditions. Narrative entrepreneurs responded by reallocating effort toward frames that reliably mobilized attention and access. Over time, this selection pressure compressed narrative diversity, favoring simplified causal stories optimized for transmission and repetition. Political outcomes thus reshaped narrative supply not by correcting or resolving disagreement, but by selecting for explanations that traveled efficiently within the existing institutional configuration.

Hardening as Mode, Not Endpoint

Across all seven edges, institutional hardening manifests as a shift in how institutions respond, not whether they act. Pressure continues to move. Decisions continue to be made. The system remains functional. What changes is the locus of discretion, the tolerance for ambiguity, and the institutional appetite for interpretive risk.

This hardened configuration does not resolve disagreement. It contains it.

Why Institutional Hardening Is a Present-State Description, Not a Forecast

Institutional hardening is not introduced here as a hypothetical future trajectory or as a contingent outcome dependent on political shocks. It is a description of a system configuration that has already emerged under existing conditions. The behaviors described in this scenario are observable now because the constraints that generate them are already in place.

Legal uncertainty surrounding fiduciary duties, disclosure obligations, and permissible consideration of non-financial factors has not resolved. Political contestation over ESG has not converged toward a stable settlement. Reputational sensitivity has intensified rather than dissipated as institutions face scrutiny from multiple audiences with incompatible evaluative standards. Capital constraints remain binding, shaped by risk budgets, regulatory obligations, and asymmetric downside exposure.

As this configuration stabilizes, it suppresses interpretive learning. Reduced discretionary variance limits opportunities for institutions to test alternative interpretations, observe differential outcomes, and clarify norms through practice. New information continues to enter the system, but it encounters hardened procedural filters that favor consistency over revision. Over time, this narrows the pathways through which learning can occur, systematically trading adaptability for defensibility. The system becomes less capable of updating its interpretations even as external conditions evolve.

Given these conditions, the institutional adaptations described in this chapter do not require additional escalation to occur. They arise from the routine operation of institutions seeking defensibility under uncertainty. Procedural reinforcement, conservative interpretation, and formalization of discretion are not emergency measures. They are sustainable responses selected repeatedly because they reduce exposure within the existing system architecture.

Institutional hardening therefore describes how the system is functioning, not where it might go. Later scenarios explore how this configuration responds to additional stress. This chapter establishes the baseline from which those responses occur.

What This Configuration Explains Well—and What It Does Not

Institutional hardening explains a specific class of observed phenomena particularly well. It explains why institutions that publicly signal commitment to ESG principles nevertheless behave cautiously, inconsistently, or defensively in practice. It explains why organizations adopt extensive procedures without producing visible directional change. It explains why activity continues while ambition narrows, and why discretionary innovation gives way to standardized compliance. It also explains why institutional behavior appears coherent across sectors despite the absence of coordination. Similar procedural adaptations emerge because similar constraints are operating, not because actors share ideology or strategy.

What institutional hardening does not explain is intentional obstruction, ideological capture, or coordinated withdrawal from contested domains. The scenario does not assume that institutions are attempting to defeat ESG, protect incumbents, or resist transition. Nor does it assume that institutions are optimizing for long-term outcomes or social welfare. It explains behavior through constraint rationality, not motive alignment.

The scenario also does not resolve normative disagreement. It does not tell us whether hardened institutions are behaving well or poorly, responsibly or irresponsibly. It describes how institutions behave when disagreement persists and resolution is unavailable. This distinction matters. Readers seeking explanations grounded in ideology, conspiracy, or moral failure will find this framework incomplete by design. Readers seeking to understand why behavior stabilizes in the absence of agreement will find it clarifying.

Why Institutional Hardening Is Stable but Fragile

Institutional hardening is stable because it is self-reinforcing. Formal procedures reduce discretionary exposure. Reduced exposure lowers the cost of continued operation under uncertainty. Over time, investments in compliance systems, documentation, training, and oversight create path dependence. Reversal becomes costly even if upstream conditions soften.

Stability is reinforced by mutual visibility. As hardened practices converge, deviation becomes more salient and therefore riskier. Institutions that attempt discretionary expansion face heightened scrutiny precisely because they stand out against a backdrop of formalism. At the same time, this configuration is fragile. Its stability depends on the persistence of uncertainty rather than its resolution. Hardened systems are resilient to disagreement but poorly adapted to sharp shocks that demand rapid interpretive judgment.

Fragility appears when formal procedures encounter conditions they were not designed to address: sudden climate shocks, liquidity crises, legal discontinuities, or abrupt political realignments. In such moments, procedural defensiveness can delay response, misallocate responsibility, or amplify risk rather than contain it. This combination—durable under steady pressure, brittle under shock—is a defining feature of the configuration. It explains why the system appears inert yet tense, orderly yet vulnerable.

How Institutional Hardening Sets the Boundary Conditions for the Scenarios That Follow

The remaining scenarios in this book do not replace institutional hardening. They operate within it. Chapter 7 examines how a hardened system responds to physical shocks from climate change when interpretive flexibility is already constrained. Chapters 8–10 examine three scenarios conditioned on hypothetical outcomes of the 2028 elections. Each scenario is grounded in the institutional hardening described in this chapter and in observed and projected climate conditions over the intervening period.

In each case, the central question is not whether institutions change character, but how a hardened configuration reallocates stress. Do pressures accumulate silently? Do they surface at particular edges? Do institutions absorb shock through further formalization, or does constraint saturation produce rupture?

Institutional hardening therefore functions as the baseline condition against which variation is observed. It defines what can change quickly and what cannot. It explains why some interventions produce outsized downstream effects while others dissipate. It clarifies why disagreement persists even as behavior stabilizes.

What follows is not a sequence of alternative worlds. It is a set of conditional responses by a system that has already learned how to protect itself from uncertainty—at the cost of flexibility, interpretive ambition, and discretionary reach. The scenarios are not predictions. They are analytical tools for examining how the institutional hardening of the anti-ESG ecosystem shapes the ways in which the US is likely to address climate policy under stress.

An impossible object, built from sentences. Every joint is locally correct. The figure cannot exist. Drawn in code by Claude (Anthropic) at my direction.

It has long been de rigueur among AI skeptics and haters to talk about errors and hallucinations. “It gave me the wrong url!” “The url was dead!” “It made up a person!” “It fabricated a quote!” “It cited an article—a book—that doesn’t exist!” I even had one friend tell me it did some addition wrong. Fond of this woman though I am, I have to wonder about the prompt she gave for that.

Do these hallucinations happen? Of course they do. Too many humans are saying they do, and I don’t believe in mass hallucinations in the human race. But let’s be honest; these are trivial matters. Check the url. Get confirmation on facts and quotes from another agent or a human expert. Make sure the journal and the book exist if you’re feeling queasy. Use the calculator on your phone to check the sums. You can even use a simple search. Though, oh dear, Google now answers with AI unless you tell it not to. So much for all these academic journals insisting you disclose the use of AI.

My guess is that as AI gets better, and it is getting better very rapidly, the hallucination problem will get smaller and smaller. The real problem is that as AI gets better it gets a great deal more confident and convincing about the analysis it hands you. It can produce amazingly fluent and cogent arguments in which every fact is right and the whole thing is still wrong, in big ways and small.

It’s a tricky one, because AI is genuinely great in ways humans aren’t. It works fast, it doesn’t get sick, it doesn’t take holidays, it doesn’t go on strike (although your context window can close and you can get hit for extra fees when you blow through your subscription), it gets you up to speed on a topic very quickly, and it is amazingly interdisciplinary. Those last two have been enormously important to my own work. They have let me explore and write in domains well outside my traditional expertise, and to combine domains in ways I could never have managed alone.

Which is exactly how I learned how easy it is to fall into the trap. When I get output that combines disciplines I know with disciplines I don’t, how do I judge the quality of the argument in the parts I don’t know? One answer is to check it—the citations, the facts, and the argument itself—against other agents, through what I call a Hybrid Intelligent Team. I am always struck by how much an argument improves when Gemini Pro is set loose as an intellectual adversary on something Claude and I have put together.

The Italians Have a Word for It: Epistemia

I had been dealing with this problem for a year without knowing it had a name or a literature. It does, and the literature is barely a year old. The word is epistemia. I first heard this word from Claude and didn’t know what it meant. He explained it to me. Eating my own dog food I also did a Google search (more AI again!) and confirmed it.

It was coined by Edoardo Loru, Walter Quattrociocchi, and colleagues at Sapienza in Rome in a paper on the simulation of judgment in large language models, published in October 2025 in the Proceedings of the National Academy of Sciences which is to say, not a fringe venue. It got its full theoretical treatment two months later in “Epistemological Fault Lines Between Human and Artificial Intelligence,” by Quattrociocchi with Valerio Capraro and Matjaž Perc.

Their definition: epistemia is the structural condition in which linguistic plausibility substitutes for epistemic evaluation. Systems produce answers that are well-formed, fluent, and rhetorically convincing without instantiating any of the processes by which beliefs normally get formed, tested, and revised. You end up in possession of an answer without having traversed the cognitive labor that would have justified it.

The etymology carries the argument. Episteme is the Greek word for knowledge resting on certain and verifiable grounds, as against doxa, mere opinion. Epistemia is what you get when the grounds go missing and everything else stays in place. It is a degenerative disease of knowing, and the Italian coverage treats it as exactly that—una nuova malattia, a new illness.

Since I’m an academic I can’t but help myself from going into the weeds on this a bit. NP if you don’t want to go there with me. You can skip to the next section and you’ll be just fine!

Epistemia is not hallucination, and not automation bias. It happens even when the answer is correct. The point is not that you were misled. The point is that you received a result without going through the process that would entitle you to it. The harm is not the production of falsehoods but the structural bypassing of evaluation. That distinction is the spine of everything below.

Epistemia is an Italian word. Treccani, roughly the Italian OED, has entered epistemìa among its 2026 neologisms—the comfortable illusion of knowledge produced by interacting with generative AI, where the simulated plausibility of fluent discourse takes the place of reliable data. Quattrociocchi had put it in the Corriere della Sera the previous July, describing an epidemic of plausible texts that simulate reasoning, argument and authority with no guarantee of any foundation. In December, Luca Zorloni made the case in Wired that it should be word of the year, under a headline that reads, more or less, "we are falling ill with epistemia, the illusion of knowing things just because the AI writes them well." Sapienza held a conference on it, with the president of the communications authority and the vice-president of the Chamber of Deputies in the room. And a columnist in Lugano reports that the word is spreading almost exclusively in Italian — a claim he arrived at, he says, by asking ChatGPT and doing some searching. Which is either the perfect illustration of this entire piece or a reason not to repeat his conclusion, and I haven't decided which.

Epistemia has a proposed mechanism, and the mechanism comes with a story worth stealing. Giulia Bini and Quattrociocchi, writing in AI & Society, propose epistemic pareidolia, the tendency to project epistemic capacity onto anything producing formally correct-looking output through processes that aren’t epistemic at all. Pareidolia is the old word for seeing the Man in the Moon or a face in a piece of toast, from the Greek pará, beside or instead of, and eídōlon, image. Their opening example is a sixth grader named Benny, studied in 1973, who reliably produced correct math answers and was thought one of his teacher’s best students. Interviews revealed he had built a set of syntactically functional rules with no mathematical meaning in them at all. Benny got credit he hadn’t earned because his output looked like knowledge. Fifty years later we do it in reverse; we meet a machine producing right answers and credit it with understanding.

Which Is Why Disclosure Won’t Save Us

The fault lines paper proposes reframing disclosure requirements as epistemic transparency obligations. What matters is not that a system was used, but which epistemic functions it failed to perform along the way: whether claims were actually grounded in something, whether sources were checked, whether uncertainty was tracked rather than smoothed over, and whether the system could have declined to answer at all. That last one is the sleeper. A human researcher who finds nothing reports finding nothing. These systems have no null output. Ask and you get fluent prose every time, which is why you cannot read confidence off the response.

That is a real improvement on “I used AI.” But I don’t think it works, for three reasons. First, it requires a reader sophisticated enough to know what the disclosure means. Say you attach a note reporting that nothing in your process was positioned to notice if the frame was wrong. To make that real you would have to explain the whole process, which would often run longer than the paper. Second, nothing in such a statement is verifiable. I can tell you my grounding agent checked five claims. You cannot check that I checked — and that assumes I am being truthful, when the deeper problem is that I could be entirely sincere and still wrong, because what I would be certifying is the absence of errors I did not notice.

And third, warrant does not come from a statement attached to a document. It comes from other people, over time, and it is hard to get. Most of the disclosure conversation is a way of not talking about that, which is the subject of the next piece.

Experiencing Epistemia

I realize these last two sections have been pretty abstract so let me give you a concrete example from my own experience. I recently finished a project in a domain where my expertise was next to zero. In a single day we produced a draft report of 20 pages and 47 footnotes using a Hybrid Intelligent Team: Claude drafting, Perplexity grounding, Gemini Pro running two adversarial reviews in different assigned voices, and a clean ChatGPT instance—memory off, told nothing about how the thing was made—doing a cold editorial pass.

I’m not going to tell you what it was about. It concerned an emerging technology question in a specialized institutional domain, written for a small audience through an intermediary whose agreement I need before it goes further. I’ll write about it properly later if I can. The subject doesn’t matter for what I want to say here.

That’s worth a pause, because it introduces a distinction the second piece is built on. Shiva Rajgopal and I have been working with a separation between claims that are warrantable and claims that are warranted. Warrantable is assessable from inside the production system. Is the justification traceable, was the process reliable, could the conditions be reproduced? Warrant is different. It is conferred by a human community that reads the thing and takes it up, and no participant in producing it can perform that step. The paper I just described has not been read by the people it was written for. It is, at best, warrantable.

The Draft We Killed on Day One

After a long day of very hard work, we had a complete draft of 6,756 words and a title I liked a great deal. On the morning of the second day, I threw all of it away. What happened was that I had spent a day trying to get on top of unfamiliar material, and somewhere in there I stopped and looked at a word the draft was leaning on constantly. The word was team. The draft assumed a stable group of named people. The domain doesn’t work that way; membership rotates. Every individual sentence was fine. No sentence anywhere stated the assumption. But the paper’s central object did not exist.

I only figured this out when I stepped back from the details to look at the argument in a holistic way. What enabled me to do this wasn’t what I’d learned in writing the draft (which was a lot 🐥) but having a Ph.D. in organizational sociology from Harvard and being a tenured professor at Harvard Business School teaching organizational behavior. It was this training that made me finally figure out the problem with the way the word team was being used. So, a sliver of domain expertise did the work, from an adjacent domain, not the one the paper was about.

Shame on me that it took me a full day to figure out that the core unit of analysis was completely mis-specified. I was too busy learning the material to look at the frame. That is exactly the condition under which epistemia operates.

The draft I killed had been through a pretty rigorous HIT process. Perplexity did some research for us. Gemini Pro had run an adversarial review in an assigned voice and returned a list of objections. Claude had adjudicated them and proposed revisions to three sections. But none of this touched the underlying problem. When Claude and I were debriefing how we got so badly off track, he admitted, “I let the title drive the argument. It’s a good image and I wrote toward it, and in doing so I quietly converted the central concept into something else without noticing.” The adversarial reviewer had found symptoms and never named the cause.

The draft that looked so good to all of us would have been laughed at by anyone who knew the field. So, we started over in a fresh conversation so the abandoned text wouldn’t exert what Claude calls the “gravitational pull” from the previous work. Over the next four days there were nine more mistakes.

Mistakes Were Made

Even after clarity had been established about the definition of the unit of analysis it was hardly smooth sailing. There were nine more places where something we had been treating as established turned out to be false, overstated, or built on a category error, and the paper changed as a result. The central claim of the finished paper is not the claim we started with. Neither is its structure, its central object, or its title. Not a single one of these reversals came from Claude re-examining his own work. Every one arrived from outside — from me, from a Perplexity query, or from a reviewer positioned differently. Several arrived immediately after Claude had reviewed the same material and found nothing wrong with it.

And get this. Every source in that paper exists and says what we say it says. Checking citations would have caught precisely nothing. So, yup, there were no hallucinations at the level of fact. The problem was much bigger than that. What’s more interesting than the number is the shape they shared. Here are three of the patterns, one of which produced several reversals on its own and one of which never became a reversal because it was caught before it could.

Confident Universal Negatives

This was the dominant failure mode. Six of them across five days. What happened here is that Claude would make claims along the lines of “No institution has ever built X.” Claude claimed something didn’t exist when it did. I guess you can call this a hallucination if you want to stretch the definition, but this is something categorically different. Claude didn’t make something up. He said nothing was there when it was. Search cannot establish a universal negative. All six arrived precisely where the argument needed a gap and Claude just went ahead and said the gap exists. We adopted a rule after the second time this happened. Any “nobody has ever” becomes a Perplexity query before it becomes a premise.

But get this. Although the rule caught every subsequent instance, Claude kept doing it 😱! The second instance appeared in the same response that acknowledged the first. The fourth arrived two days after the rule was adopted. The rule was a smoke detector, not a fire suppressant. Remove it and the rate returns to baseline immediately, because it never left. A control that reliably catches a failure without reducing its incidence has a specific property. Its value depends entirely on somebody continuing to apply it, forever, with no accumulated improvement to fall back on. And that’s the role of what I call the Human Coherence Anchor.

An Error that Came in Through the Checker

This is a more subtle issue, and I guess you can call it a hallucination if it makes you feel better. It was a mischaracterization of a federal statute we got from Perplexity. The citations were absolutely correct, but the interpretation wasn’t quite right, although it wasn’t completely out to lunch either. Think about it. Humans have different interpretations of facts like laws and statutes all the time. This is where domain expertise is important and in this case I didn’t have it. As a result what Perplexity told us went into two sections and survived an adversarial review, a full rewrite, and my own complete read. Nobody re-examined it, because it had all the surface features of verified material—citations, document numbers, and a stated method—and those features are exactly what we had trained ourselves to accept.

Our rule above covered claims Claude generated. It said nothing about claims Claude received. A checking architecture built around distrusting your own output will not catch errors entering through the channel you built to do the checking. Another good argument for a HIT but even then mistakes can remain. For all I know—in fact I’m sure—problems like this remain in the final document. What will catch them is when humans with the necessary domain expertise read it. And even then, there’s a good chance—I’d predict a certainty—that there will be disagreements among the human experts.

A Category Collapse Across Six Passages

This was the most difficult one of all to catch, and it shows what the Human Coherence Anchor is actually for. Not catching this error—I missed it completely—but having built a team with someone in it who could. What happened was a subtle morphing of sentence claims which resulted in a paragraph which was wrong in the aggregate. “Who holds the record” quietly became “who updates it,” then “what it remembers,” then “who may use it.” No sentence ever asserted these were the same thing.

The error lived in the relationships between sentences, which is why both Gemini Pro runs read those passages and flagged nothing. Each sentence was plausible. The paragraph was not. It took the clean ChatGPT instance, briefed to test whether each step followed from the one before, to see it. And that is the anchor’s real job: knowing that a project like this needs a reader who has been told nothing about it, and putting one in the room before you need him. This is an example of epistemia operating on me—not on some undergraduate pasting prompts into a chatbot, but on a project with five reviewers, a written rule, and a human actively hunting for this specific failure.

Dealing with Epistemia

Along with hallucinations, another main theme in the prevailing narrative about AI is that it takes out the routine work. This is probably true and an important issue to deal with but that’s not what I’m focused on. I’m looking at knowledge work where the job hasn’t been eliminated but AI is dramatically changing how it can be done. Shiva Rajgopal and I have written about this in “The Artisanal Paper Is Dead.” AI dramatically compresses the production process, but it does not eliminate the time needed to check the output you got from AI.

This has two important implications. The first is that AI can substantially increase output, including making this output more interdisciplinary. The second is that human time shifts from production to validation since there is more output to validate. But proper validation requires domain expertise, strong analytical skills, and the ability to think rigorously. The worry here is that using AI in the production process will weaken the muscle memory needed for validation or, in younger generations, preventing the development of muscle memory in the first place.

I’m 75 and have been developing muscle memory (and some real muscle from my weightlifting 🏋️‍♀️) for some 60 years or so I’m not all that worried about this. If anything, I’m finding I need to develop new muscles as my work becomes more interdisciplinary and new exercises always make you sore until you’re stronger. This is a good thing. I should also note that when I’m outside my domains of expertise I engage deeply with humans who have this expertise.

The process I’m describing makes output warrantable. The justification is traceable, the process is reliable, and the thing won’t be totally embarrassing when read by human experts. But there’s a long way to get from “not embarrassing” to “this is a decent piece of work” and that’s the subject of my next piece.

This chapter defines the methodological constants that structure the remainder of this book. These constants are not empirical claims offered for validation, nor normative positions offered for agreement. They are analytical commitments. Their function is to stabilize interpretation, discipline causal reasoning, and ensure that disagreement is located at the level of assumptions rather than displaced onto outcomes. The analysis that follows in subsequent chapters is conditional on these constants. Holding a condition constant is not a claim about its accuracy, but a decision about what is not being explained here.

Readers who disagree with later conclusions are therefore invited to return to this chapter and identify which assumptions they reject. Disagreement that does not pass through this step risks becoming incoherent, arguing over effects while leaving causes implicit or contesting outcomes while accepting the premises that generate them.

The constants are organized into seven domains: system architecture, institutional behavior, legal uncertainty, capital constraints, firm adaptation, macro conditions, and the nature of disagreement itself. Each domain is held constant for analytical purposes. None is defended as descriptively complete or universally true. The purpose is not to foreclose debate, but to give it a clear starting point.

System Architecture Is Treated as Fixed

The analytical system used in this book is assumed as given. Its boundaries, components, and internal structure are held constant throughout the analysis. The system is not presented as a full representation of the political economy surrounding ESG, nor as the only valid way to describe it. It is presented as a defined analytical construct within which causal relationships can be examined consistently.

The system consists of seven interacting edges. Each edge represents a structured transmission pathway through which incentives, constraints, and signals propagate. The edges are analytically distinct but empirically interdependent. No edge is treated as derivative of another, and none is granted default causal priority. Holding the system architecture constant serves three important methodological purposes.

· First, it constrains explanation. Without a fixed architecture, analysis can expand indefinitely, incorporating additional actors, motives, or domains whenever outcomes appear puzzling. Such expansion may feel explanatory, but it weakens causal discipline by allowing the analytical frame to shift in response to results.

· Second, it enables comparability. A stable architecture allows different conditions, scenarios, or parameter values to be examined within the same structural frame. Differences in outcomes can then be attributed to differences in inputs or interactions rather than to changes in the analytical model itself.

· Third, it prevents retrospective redesign. Analytical systems are often modified after outcomes are observed, implicitly tailoring structure to fit results. By holding the architecture fixed, this analysis resists that tendency.

Disagreement with the system architecture is possible and legitimate. However, such disagreement must be explicit. Arguments that critique conclusions derived from this system while implicitly relying on a different architecture are analytically misaligned. The analysis that follows is conditional on the system as defined, not on alternative frameworks that may be preferred.

Institutional Behavior Is Assumed to Be Constraint-Rational

Institutions in this analysis are assumed to behave rationally within constraints ( i.e., bounded rationality). Rationality here is not defined as global optimization or perfect foresight. It refers to behavior that is internally coherent given institutional mandates, incentive structures, information limitations, and procedural constraints.

Institutions include firms, asset managers, asset owners, regulators, legislators, courts, and intermediaries. They are treated as organizations rather than as aggregations of individual beliefs or ideologies. Their actions are interpreted through formal responsibilities, risk exposure, and reputational considerations rather than through personal conviction. Four implications follow from this assumption.

· First, institutional behavior is not explained primarily by ideology. Institutions may use ideological language, but such language is treated as communicative or justificatory unless it demonstrably alters incentives or constraints.

· Second, bad faith is not a default explanatory mechanism. The analysis does not rely on hidden motives, deception, or coordinated intent to explain observed patterns. When outcomes appear inconsistent with stated objectives, the explanation is sought in constraint misalignment rather than duplicity.

· Third, institutions respond to incentives even when doing so produces outcomes that are unintended or undesirable from their own perspective. Rational behavior at the institutional level can generate systemic effects that no participant explicitly seeks.

· Fourth, institutional change is incremental. Adaptation typically occurs through policy adjustments, compliance strategies, and procedural revisions rather than abrupt strategic reversals. Sudden shifts are interpreted as responses to external shocks rather than spontaneous changes in preference.

Holding institutional rationality constant narrows the explanatory space. It excludes accounts that rely primarily on moral failure, cultural capture, or conspiratorial coordination. This exclusion is methodological rather than normative. Such phenomena may exist, but they are not required to explain the dynamics examined here. Readers who attribute outcomes mainly to ideology or bad faith are operating with a different explanatory baseline. The analysis that follows does not attempt to adjudicate that baseline.

Legal Uncertainty Is Treated as Structural and Persistent

Legal uncertainty is assumed to be a durable feature of the environment rather than a temporary deviation awaiting resolution. This uncertainty arises from overlapping jurisdictions, evolving statutory interpretation, regulatory discretion, and the inherent lag between innovation and adjudication. In the context examined here, legal uncertainty manifests as ambiguity around fiduciary obligations, disclosure standards, permissible consideration of non-financial factors, and enforcement risk. These uncertainties are not assumed to converge quickly toward clarity. Four constants follow from this assumption.

· First, institutions operate under conditions of incomplete legal certainty. Decisions are made based on probabilistic assessments of risk rather than definitive legal guidance.

· Second, legal risk is asymmetric. The downside of adverse judgment or enforcement often exceeds the upside of permissive interpretation. This asymmetry biases behavior toward caution even when permissibility is plausible.

· Third, legal uncertainty interacts with reputational exposure. Institutions frequently treat reputational risk as a proxy for legal risk, especially when formal standards are unsettled.

· Fourth, institutions operating across jurisdictions tend to standardize behavior toward the most restrictive or uncertain regime rather than aggressively arbitraging differences.

Holding legal uncertainty constant avoids explanations that presume clear legal resolution as a prerequisite for behavior. It also avoids attributing conservative or heterogeneous responses to confusion or incompetence. Such responses are instead treated as rational under persistent uncertainty. Analyses that assume imminent legal clarity or decisive settlement will diverge from the framework used here. That divergence should be recognized at the level of assumptions.

Capital Is Assumed to Be Constrained

Capital is treated as constrained rather than frictionless. This constraint applies not only in aggregate but also at the level of deployable capital given institutional mandates, liquidity requirements, risk limits, and regulatory obligations. Capital constraints operate along four dimensions.

· First, capital is not perfectly substitutable. Capital allocated under specific mandates cannot be freely redeployed without violating contractual, fiduciary, or regulatory constraints.

· Second, time horizons matter. Short-term liquidity needs and performance evaluation cycles constrain long-term positioning even when long-term returns appear attractive.

· Third, risk budgets are finite. Institutions allocate risk capacity across competing exposures. An increase in perceived risk in one domain reduces capacity elsewhere.

· Fourth, transaction costs—financial, operational, and reputational—affect decision thresholds and slow reallocation.

Holding capital constraints constant prevents explanations that rely on unlimited capital mobility or frictionless adjustment. It also avoids attributing observed outcomes to preference or ideology when constraint suffices. Importantly, capital constraint does not imply macroeconomic scarcity. Even in periods of abundant liquidity, deployable capital remains bounded by institutional structure. Readers who assume near-perfect mobility or infinite substitutability are working from a different premise than the one adopted here.

Firm Adaptation Is Strategic but Bounded

Firms are assumed to adapt strategically to their environments. They monitor regulatory signals, investor expectations, consumer behavior, and competitive dynamics, and they adjust disclosures, operations, and investment accordingly. At the same time, firms are not assumed to be omniscient or perfectly anticipatory. Four constants define firm adaptation.

· First, adaptation is constrained by legacy assets, organizational structure, and sunk costs. Firms cannot pivot without incurring friction.

· Second, adaptation is probabilistic. Firms respond to perceived distributions of outcomes rather than to certain futures.

· Third, signaling matters. Firms adapt not only to realized constraints but also to anticipated scrutiny, which can produce preemptive or hedging behavior.

· Fourth, adaptation occurs under uncertainty about peer behavior. Competitive dynamics influence both the pace and direction of adjustment.

Holding firm adaptation constant avoids two symmetrical errors: treating firms as passive recipients of external pressure and treating them as hyper-rational actors capable of engineering outcomes with precision. The analysis assumes responsiveness without omnipotence. System-level outcomes are therefore understood as emergent results of aggregated adaptation rather than as products of coordinated design.

Macro Conditions Are Treated as Exogenous

Macroeconomic conditions—growth, inflation, interest rates, geopolitical shocks—are treated as exogenous to the analytical system. They affect the system but are not explained by it. This assumption has three implications.

· First, macro volatility is treated as background variation unless it directly alters institutional constraints or incentives within the system.

· Second, cyclical effects are not attributed to system dynamics unless they propagate through defined edges.

· Third, macro conditions are held constant across analytical exercises to isolate system behavior.

Holding macro conditions constant reduces confounding and improves interpretability. It allows the analysis to focus on interactions within the system rather than on external shocks. This choice limits the scope of the analysis by design. Readers interested primarily in macro-level causation will find the framework intentionally narrow.

Disagreement Is Treated as Upstream and Structural

Disagreement is treated as an object of analysis rather than as an obstacle to be overcome. The analysis assumes that persistent disagreement arises primarily from differences in assumptions rather than from differences in values or information alone. Three constants define this treatment.

· First, disagreement is assumed to be sincere. Participants are presumed to interpret the system differently rather than to argue in bad faith.

· Second, disagreement is structural. It persists because analytical baselines differ, not because misunderstanding can be fully eliminated.

· Third, disagreement is not resolved by observing outcomes. The same outcome can be interpreted differently depending on prior assumptions.

Holding disagreement constant shifts the analytical task. Rather than adjudicating which position is correct, the analysis seeks to clarify how different conclusions follow from different starting points. This approach does not neutralize disagreement. It locates it.

Analytical Consequences of Holding These Constants Fixed

Holding these constants fixed serves three methodological functions.

· First, it improves interpretability. Conclusions can be traced back to defined assumptions rather than inferred motives.

· Second, it sharpens disagreement. Readers are directed to contest premises rather than results.

· Third, it reduces false causal attribution. Explanations are constrained to operate within defined bounds rather than expanding ad hoc.

These benefits come at a cost. Certain explanations are excluded, and complexity is compressed. That cost is accepted in exchange for analytical clarity.

How to Read What Follows

Subsequent analysis proceeds within the bounds defined here. Interpret later conclusions by first checking whether they follow from the constants established in this chapter. If a conclusion appears implausible, identify which constant would need to change for that conclusion not to follow. If an implication feels objectionable, determine whether the objection targets the analysis or the assumptions that structure it.

This chapter does not ask for agreement. It establishes the ground on which disagreement will occur. What follows depends on what is held constant here.