Subscribe now

On May 29, the Office of Management and Budget published a proposed rule that would fundamentally reshape how the federal government funds scientific research. The rule covers dozens of agencies — NSF, NIH, NASA, DOE, USDA, and more — in 412 pages of proposed revisions to 2 CFR Part 200, the government-wide grants management framework. Public comments are due July 13, 2026 at docket OMB-2026-0034.

This post evaluates the proposed rule through the framework of my 2007 book The Honest Broker: Making Sense of Science in Policy and Politics (Cambridge University Press, 2007). I argued that science supports democratic deliberation and decision making when it remains open, contestable, and simultaneously engaged with politics yet not captured by politics. Science becomes pathologically politicized when it becomes a tool for short-circuiting democratic decision making, such as when political actors embed their preferred conclusions into the machinery of expert authority via stealth advocacy.

For eighty years, federal science funding has rested on a social contract rooted in Vannevar Bush’s 1945 Science: The Endless Frontier: the public provides resources and a high degree of autonomy, and in return scientists produce knowledge that ultimately serves the public interest. That contract has been under strain for some time now, as often discussed here at THB.

Alvin Weinberg gave that contract a useful characterization in “Criteria for Scientific Choice” (Minerva, 1963): “science for policy” — research that informs public decisions, and “policy for science” — government stewardship of the research enterprise.

The proposed OMB rule would redefine both halves of that contract in explicitly political and partisan terms, collapsing Weinberg’s productive tension into one-way, top-down executive control. The result would not be better science policy — or even science policy at all. Instead, the new rule offers pathologically politicized science dressed up in regulatory language.

Federal science funding has real problems — replication failures, lack of rigor, and ideological capture in some programs. Even so, the proposals of the OMB rule are far worse than the problems they claim to address.

This week, Science editor-in-chief Holden Thorp wrote that the Trump administration “seems as determined as ever to mortally wound the nation’s scientific enterprise” and issued a call to action at his Substack. I agree with Thorp that the scientific community needs to mobilize a response — but how it mobilizes matters enormously, and if done poorly, could even make things worse. More on that later in this post.

First, let’s take a look at five problems with the proposed new OMB rule.

Problem #1: Political Appointees Now Formally Outrank Peer Review

“[S]enior appointees (or their designee) must not ministerially ratify or routinely defer to the recommendations of others, but must instead use their independent judgment when evaluating Federal award proposals.” — §200.205(c)

This single sentence proposes a consequential structural change in the proposed rule: all discretionary research awards must “demonstrably advance the President’s policy priorities.” Peer review in awarding extramural research grants has always been advisory — as anyone who has received or reviewed a grant well knows. However, the decisions of panels and program officers have never been subject to political review for conformance to a political agenda.

That by itself represents a shredding of the social contract for science.

In The Honest Broker, I identified the “science arbiter” role — the expert who provides authoritative technical answers to questions posed by decision makers so policy choices remain visible and contestable. Policy makers pose the questions — enlisting expertise as a part of the political process, and experts offer answers (complete with uncertainties and contestation) — keeping science apart from political meddling.

Once political officials override experts, that division of responsibility collapses. What remains is what Sheila Jasanoff (1990) called “regulatory science” that is captured by its principals. Russell Vought’s own Project 2025 chapter described OMB as the “keeper of ‘commander’s intent’” — this OMB rule operationalizes that vision.

Experts are reduced to providing the “science” that political leaders prefer to have. And that is not science at all, but pathologically politicized science.

Problem #2: “Gold Standard Science” Is a Political Label, Not a Scientific Standard

“[A]pplicants should commit to complying with administration policies, procedures, and guidance respecting Gold Standard Science.” — §200.205(b)(5)

“Gold Standard Science” appears five times in the regulatory text but receives no definition in 412 pages. Its origin traces to Executive Order 14303 (”Restoring Gold Standard Science,” May 2025), which is equally vague. An undefined standard is ad hoc, inconsistent, without accountability — a slogan, not policy.

For instance, the rule prohibits funding for activities that deny “the sex binary in humans or the notion that sex is a chosen or mutable characteristic” — embedding into federal funding criteria a specific position on an empirical question that is perfectly reasonable for experts to discuss and debate. Imagine a Democratic administration doing the mirror image — prohibiting funding for activities that reflect a view that “the sex binary is absolute.”

Science is science because researchers have freedom to research, debate, and discuss controversial subjects. In science it is OK to be wrong. Science exists to sort out competing knowledge claims, not to reflect a catechism built upon dogma.

In The Honest Broker, I described the “stealth issue advocate” — the actor who encodes contested value and empirical conclusions into supposedly neutral governance frameworks, making political choices invisible and removing them from democratic accountability. This is not a new pathology: Jasanoff (1987) documented it in EPA and FDA regulatory science. When contested empirical claims enter grant criteria, science stops serving democratic deliberation and starts foreclosing it.

Problem #3: The Rule Builds Regulatory Architecture on Partisan Politics

“Federal programs were frequently designed between 2021 and 2024 to include preferences and selection criteria aimed at advancing identity-based DEI policies.” — Preamble §A.1 (citing Heritage Foundation Backgrounder No. 3710 and Heritage Foundation report, Apr. 18, 2024)

Federal rulemaking carries legal and democratic weight because it builds on a foundation that historically has been viewed to be authoritative, legitimate, and salient — such as peer-reviewed literature, government audits, and Inspector General reports. That doesn’t mean that rules are always beneficial or based on science that might later be revisited — only that democratic legitimacy has been served.

The rule does cite legitimate sources — a 2023 GAO report, several IG findings. But the ideological narrative that justifies the rule’s most consequential changes rests on partisan advocacy documents, such as Russell Vought’s Project 2025 chapter. The significant literature on science and technology policy and science and technology studies has no presence.

A rewriting of the social contract that has governed U.S. research and development for more than 80 years may well be a good idea. I have argued for thirty years that the existing social contract for science needs updating. A serious rewriting might well be warranted. But to succeed, such a project must build upon broad democratic legitimacy — not narrow partisan authority.

Problem #4: The Foreign Collaboration Ban Would Compromise U.S. Science

“Federal funds may not be obligated or expended by a recipient or subrecipient to support a bilateral or multilateral collaboration, agreement, program, or activity with a covered foreign country or covered foreign entity.” — §200.220(a)

The proposed rule would prohibit federal funds — including indirect costs — from supporting any bilateral activity with a “covered foreign country.” The rule provides no scientific impact analysis, no estimate of affected grants, and no mechanism for scientific review of exceptions. An agency head may grant exceptions upon a national-interest determination requiring no expert input and no documentation standard.

Modern science is structurally international. A 2023 Nature analysis found that internationally co-authored papers draw substantially higher citations. The NSB’s 2024 Science and Engineering Indicators documents that U.S. researchers co-author with Chinese counterparts more than with researchers from any other country.

As Caroline Wagner showed in The New Invisible College (Brookings, 2008), modern science operates as a global network that centralized national control cannot replicate or replace. Preventing technology transfer to adversaries is a legitimate goal. A blanket prohibition with no scientific triage, no impact analysis, and no expert review process is not a serious policy instrument.

Problem #5: The Rule Substitutes Partisan Acceptability for Democratic Accountability

“[S]enior appointees (or their designee) must not ministerially ratify or routinely defer to the recommendations of others, but must instead use their independent judgment when evaluating Federal award proposals.” — §200.205(c)

The rule creates a comprehensive new accountability architecture for grantees: pre-issuance review, mandatory payment justifications, expanded risk criteria, E-Verify requirements, subaward reporting, and new compliance conditions. That accountability structure ends with political appointees making final judgments.

Former NIH Director Elias Zerhouni — appointed by George W. Bush — told Science magazine that the rule’s structure alarmed him. In that same article, Neal Lane, who served as science advisor to President Bill Clinton, agreed, citing the case of NSF:

“NSF makes grant decisions based on the deep knowledge of experts in the field, not on whether it meets an ideological agenda. That’s what merit review is all about. Replacing it with top-down decision-making will destroy that process and result in bad science being funded.”

There are many aspects of federal R&D that might be improved — giving political appointees ultimate power to approve or deny who or what is funded would not be an improvement.

What Should Happen Next

The comment period closes July 13, 2026. Substantive comments addressing specific regulatory text carry more legal weight than general objections. Here is a short action agenda with specific actors named.

1. OMB: Withdraw or substantially revise the proposed rule

Merit review criteria must rest on peer-reviewed scientific standards, not on undefined political labels or contested empirical claims written into grant conditions. OMB should develop specific, measurable quality standards through consultation with NSF, NIH, and the scientific community — not sourced from partisan manifestos. This is of course unlikely to happen.

2. Congress: Exercise oversight now, and make it bipartisan

The House Science Committee and Senate Commerce Committee should hold hearings before this rule takes effect on October 1, 2026. Crucially — Republicans should note the fundamental asymmetry: partisan control of federal science governance via the executive branch would not disappear when Democrats win the next (or some other) election.

That means that every elected official representing a major research university, medical center, or national laboratory has a direct constituent interest in improving federal science policy in a way that benefits all Americans, regardless who happens to be in office.

Congress should in parallel advance scientific integrity legislation. In my 2019 testimony before the House Science Committee, I argued that scientific integrity obligations fall on appointed officials, not only on scientists, and that such protections require bipartisan support to survive changes of administration.

Rep. Paul Tonko (D-NY) and Rep. Brian Fitzpatrick (R-PA) co-introduced the Scientific Integrity Act (H.R. 1106) in the 119th Congress — with 111 Democratic and 1 Republican cosponsor beyond the two sponsors. As I discussed in “Trump vs. Biden on Science Integrity” at THB last August, that ratio remains a problem: scientific integrity protections mean nothing if they are not bipartisan. Congress should engage federal science policy in a bipartisan manner, starting now.

3. OMB: Replace the rule’s international provisions with a targeted national security review process

Replace the blanket prohibition on foreign collaborations with a risk-stratified framework by research type, institution, and collaboration structure. Federal agencies and units — like JASON, the NSF Inspector General, and the intelligence community — have the tools. Use them. Fix and improve processes — don’t just break them.

A Warning: Don’t Try to Fight Partisan Fire with more Partisan Fire

Leaders in the scientific community should take on the proposed new rule, but how they take it on makes a huge difference.

Casting debate over the rule as science-versus-Trump would hand the rule’s authors the argument they want. The preamble argues at length that federal science functioned as an arm of the political left for four years — an argument with enough traction that a scientists-versus-MAGA frame simply confirms the rule’s own diagnosis.

If the scientific community accepts or reinforces that framing it would tell the public: yes, this is a culture war, and we scientists are on one side and half of America is on the other side. The scientific community should not walk into that trap.

As well, every Republican who might cheer on this rule today should understanding that they are handing a a powerful political weapon to the next Democratic administration. Permanent executive control of science funding serves no one’s long-term interest.

That requires leaders in the scientific community to do several things that they would surely find uncomfortable: acknowledge that some programs became ideologically captured during the Biden years (and even earlier); build a coalition that explicitly includes Republican scientists, Republican university presidents, and Republican members of Congress with major research institutions in their districts; and frame this as a defense of American national capacity, not framed as a defense of a Democratically-aligned scientific establishment against the MAGA Republicans.

The social contract that has sustained federal investment in research for eighty years — producing along with the private sector modern medicine, the internet, GPS, and the foundations of American agriculture — does not belong to either political party. This proposed rule treats that contract as a political instrument of the current executive. That is a bad idea whether the president is a Republican or a Democract.

Science is for all Americans. This rule is anti-American.

Comments welcome!

Leave a comment

Before you go — Please click that “❤️ Like” button. More likes mean that THB rises in the Substack algorithm and gets in front of more readers. Thanks!

THB is a community — honest brokering is a group effort! Please share, subscribe, and if already subscribed, consider an upgrade to support this growing community!

Subscribe now

Subscribe now

When I decided to leave academia, a few years ago, I was 55, which was exactly the median age of U.S. tenure-track faculty. In theory, that means I was mid-career, and in practice that seems about right! Even so, leaving academia gave me a chance to look back at my work over the past 30 years, since getting my PhD in 1994.

I set out to write a post about my five favorite papers I’ve authored or co-authored over that time. However, it turns out that I have more than five favorites! So today I am sharing five of my favorite papers, and will do another post like this down the road.

Why these five papers?

These are certainly not my most read or most cited papers.

I have applied two criteria:

• First, these papers made what I see as really novel arguments, and

• Second, these are ones that I can still recall the intense feeling of learning and clarity while writing. That is a hard feeling to explain.

The closest thing I can compare it to is the feeling of being in “the zone” when playing a competitive sport.

When asked about how one gets into “the zone,” Michael Jordan explained:

"I don't think anyone knows. I think it happens when you put forth the work ethics. You know, and then next thing you know, you are achieving what you always worked hard to get, but you can't get it without putting forth the work."

That sounds right to me.

Today, I share these papers (with ungated PDFs) from the zone and and also share the inside backstory to each.

Note: Don’t forget the huge THB discounts for students and military/government.

1. Pielke Jr., R. 1999. Who Decides? Forecasts and Responsibilities. Applied Behavioral Science Review, 7:83-101.

Backstory

In spring 1997, there was massive flooding along the Red River of the North which flows south to north between North Dakota and Minnesota. Fargo and Grand Forks experienced massive destruction. In the weeks following, I was invited by the U.S. National Weather Service to serve on its disaster survey team.

The main focus of our inquiry was how it could be that a flood that was well-forecasted, months in advance, resulted in emergency evacuations in the middle of the night in communities seemingly unprepared, despite the advance warnings.

Our team went to North Dakota and Minnesota, where we had a chance to interview forecasters, elected officials, emergency managers, and members of the public. The experience for me was career changing — driving home the fact that even the very best science doesn’t matter unless it is effectively used in decision making.

This paper was the result of my efforts to make sense of how an accurate forecast could be coupled with disastrous outcomes. You can see our survey team’s official NWS report here. Sometime, I’ll do a full public post on the incredible experience.

Postscript: I’m not 100% sure, but I am pretty confident that it was this paper that brought my work to the attention of Nate Silver who discussed it in his book, The Signal and the Noise: Why So Many Predictions Fail, But Some Don’t.

Nate then hired me at 538 in 2014 to write on science (and sports) and then quickly dropped me after a Twitter cancellation campaign by activist climate scientists, journalists, and randos. But that is another story.

Refer a friend

2. Pielke, R. A. 2012. Post-normal science in a German landscape. Nature and Culture, 7:196-212.

Backstory

This paper reflects my exploration of how we make decisions in the face of ignorance, which I characterize in the paper as the “neglected sibling of uncertainty.” I have been fortunate over the decades to have spent a lot of time in Germany, and to observe its efforts to decarbonize its economy in the face of incredibly complex politics — characterized as the energiewende (energy turnaround).

I decided the write the paper after Matthias Groß, a German colleague, took me on a tour of the human-made lakes south of Leipzig (see the contemporary video above), which were created from open-pit lignite (coal) mines. I can vividly recall the beautiful setting, with wind turbines slowly spinning on the horizon, and trying to reconcile the challenges of economic development, energy security, environmental policy (including climate), and incredibly complex politics.

I remember thinking — no one knows how to do this, and yet we must. This paper was the result of my thinking through these issues. Ignorance is both pretty common and a surmountable obstacle to effective decision making. The paper concludes:

Sitting there in the gentle breeze in the transformed landscape enjoying my lunch and conversation with Matthias Groß I could almost imagine that Germany had figured out something really important about innovation, democracy, and building a better, shared future. Almost.

Refer a friend

3. Pielke Jr, R., & Linnér, B. O. 2019. From Green Revolution to Green Evolution: A critique of the political myth of averted famine. Minerva, 57:265-291.

Backstory

This was an extremely fun paper to research and write, and is co-authored with my friend and colleague Björn-Ola Linnér, who has written an excellent book on Malthusianism. This paper had a much longer gestation than any paper I’ve written, having been started in 2011 during an extended stay in Linköping, Sweden, but not published until 2019.

Not only was the research quite intensive, but during the writing my career was interrupted by the 538 cancellation campaign, attack by the Obama White House, and congressional investigation.

Ironically, this paper marked an effort to start diversifying my work beyond climate in the midst of relentless attacks and efforts to derail my career by academic colleagues.

When I did step away from climate research for several years, I had an opportunity to finalize this paper, which really helped me to better understand the population scares of the 1960s and 1970s, as well as the so-called “Green Revolution.”

I learned that the stories we tell in academia — simply because we have heard them uncritically so many times — are not always grounded in reality.

In this case, the “Green Revolution” is far more complicated than Norman Borlaug (a great man to be sure) saving the world from famine. The detail in this paper about Indira Ghandi and Lyndon Johnson conspiring to fool the world is a doozy.

Refer a friend

4. Pape, M., & R. Pielke Jr. 2019. Science, sport, sex, and the case of Caster Semenya. Issues in Science and technology, 36:56-63.

Backstory

In February, 2019, I served as a pro bono expert witness in the arbitration case of Caster Semenya vs. the International Association of Athletics Federations (IAAF), in Lausanne, Switzerland. I came to be a witness based on research (with Ross Tucker and Erik Boye) that had identified egregious data errors in a paper the IAAF was using to ban athletes from competing at the distances that Semenya ran (and only those distances).

Those errors were obvious, admitted to by the IAAF, and consequential — yet the journal that published them refused to retract the paper, but I digress.

Later in 2019, I was invited by the government of France to participate in a symposium following up the arbitration (which Semenya lost in a 2-1 decision), which included experts from both sides of the case. Imagine my surprise when I saw the lead IAAF expert presenting a slide (shown above) with more false information.

Here we go again, I thought.

At that symposium I met Madeleine Pape, who had just completed a PhD in the sociology of science (and happens to be an Olympian who competed against Semenya). We decided to collaborate on what would become this paper, which explores how science is created in support of a political agenda — which this case demonstrates clearly.

Specifically, in order to study differences between men and women, researchers need to first classify men and women. If this research is then used to justify how to classify men and women, we arrive at complete circularity — it is not the research that dictates the classification, it is pre-research classification. More recently, I’ve come to call this dynamic tactical science.

Postscript: The paper that we identified the errors in and which was used to justify gender regulations underwent a “massive correction” (the last sentence in that Retraction Watch article is a zinger).

Refer a friend

5. Pielke, R. (2018). Opening up the climate policy envelope. Issues in Science and Technology, 34(4), 30-36.

Backstory

This paper marked my decision to return to climate research again in 2018, after taking several years away. I was asked to give a talk in Tokyo and this paper was the result.

It turns out that the cancellation campaign against me was actually fortunate — It gave me motivation to develop new emphases in my research, allowing me to explore new areas and meet many wonderful new people.

It also allowed me to take a fresh look at climate science and policy after some time away. One important motivation for the fresh look was the work of Justin Ritchie (who since has become a valued collaborator) which led to many “Aha!” moments.

I also learned that I can’t be cancelled. THB exists because of these experiences.

With hindsight, this paper provided a roadmap for my extensive work (especially with Ritchie and Matt Burgess) on climate scenarios that THB readers have become very familiar with over the years. I write in this paper:

Scenarios are essential because to move into the future intentionally we need some expectation of how actions and outcomes may be related. But scenarios may become captured by assumptions and beliefs about how the world does or should work, and thus can limit our vision of possible futures, and make us vulnerable to surprises.

Postscript: It is no surprise that this is the second paper of the five that appeared in Issues in Science and Technology. That is no doubt thanks to the former editor of IST, Dan Sarewitz, whose editing makes everything better. Thanks Dan!

Comments welcomed!

Leave a comment

Before you go: To support THB, please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

Subscribe now

In May I continued to discuss the official retirement of RCP8.5. I also shared our new research on demographic forecasting, took a look at institutional partisan advocacy, and highlighted John Stossel’s excellent video on my departure from academia. Below are links and capsule summaries.

In the coming weeks I’ll be posting the next THB Deep Dive looking at ENSO and its relationship with U.S. hurricanes, global disaster losses, and Colorado River streamflow. I’ll also be sharing a new preprint (with Jessica Weinkle) that updates our time series of normalized continental U.S. hurricane losses. June will be fun — but before all that, let’s look back at May.

“The World’s Most Important Science Advisory Committee” mapped the governance architecture of an obscure science committee that controls which scenarios enter climate modeling. Few people know anything about this committee, who is on it, what organizations it reports to, or who funds it. The governance of climate-research-for-policy is problematic, to say the least.

In the weeks following RCP8.5 RIP there was near-silence from mainstream U.S. media outlets. That changed in a hurry when Donald Trump posted about it near the end of the month. But before Trump’s post, the biggest story in climate science in years received almost no coverage.

This post summarized the 2021 paper — by me and Justin Ritchie — documenting how RCP8.5 (and more generally the RCPs and SSPs) emerged from a suite of mistakes and confusion setting the stage for the scenario mess that the climate community now finds itself in. These problems are not solved. Tip for new readers — start here for background on climate scenarios which I’ve been discussing here at THB since the beginning.

Spinmeisters have been hard at work trying to invent an alternative history: Specifically, that policy progress — not flawed theory and faulty assumptions — were the reason for the RCP8.5 retirement. Improving climate research and its use in decision making depends in no small part in accurate and honest understandings of how we got to today. The evidence in this post is irrefutable — That is why no one has even tried to refute it.

Trump’s post on RCP8.5 got one thing correct — RCP8.5 was WRONG! But pretty much everything else in his post was false. The first wave of U.S. media coverage followed Trump’s lead by reporting on RCP8.5 with many false claims. This post documents some of those false assertions. I do want to acknowledge that later coverage was better, and the New York Times, in particular, had a very good second story on RCP8.5.

This post looks at the increasingly partisan posture that major scientific institutions adopted after 2016, and takes a close look at new research that deepens our understanding of the consequences of Nature’s presidential endorsements in 2020 and 2024. Not surprisingly, the endorsements didn’t have any effect on how people voted, but they did lead to a loss of public trust in science — not just among those on the political right, but among moderates as well. Leaders of science institutions should pay close attention.

This post introduced a new preprint co-authored with Samuele Lo Piano, Marta Kuc-Czarnecka, and Andrea Saltelli. The central finding: model choice — not parameter uncertainty — accounts for almost all of the variance in long-run population projections. The model you pick determines the answer far more than any technical assumption you make within it. That has important implications for those who use population projections (including climate modelers), and we suggestion some guidelines for their use.

John Stossel’s short account of some of my experiences that led me ultimately to choose to leave academia is very well done. Newer readers wanting background on how I got to where I am today, and why I’d do it all over again — start with this one.

Comments, suggestions, requests — All welcome!

Leave a comment

Before you go: If you like these month-in-review posts, please let me know by clicking that “❤️ Like” button on today’s post. Thanks!

THB exists because of its subscribers — Thank you!

Subscribe now

Subscribe now

We have just published a new preprint — What demographic prediction can and cannot achieve — that argues that the choice of model drives nearly all the variance in long-run demographic projections, not data quality, not parameter uncertainty. Different modeling frameworks can lead to projected global population for 2075 varying by tens of billions. Fertility assumptions account for roughly 2% of output variance. That means that the numbers we get from any single demographic model tell us more about that model than about the future.

In the preprint, my colleagues Samuele Lo Piano (Gdansk University of Technology), Marta Kuc-Czarnecka (Gdansk), and Andrea Saltelli (Universitat Pompeu Fabra / University of Bergen) and I argue why we should not treat any single demographic trajectory as authoritative. Today’s post summarizes our argument.

Before jumping in, let’s define several key terms:

• Demography “is the study of the size, territorial distribution, and composition of population, changes therein, and the components of such changes, which may be identified as natality, mortality, territorial movement (migration),

  and social mobility (change of status).”

• Demographic models: “refer to all mathematical, statistical, forecast, and microsimulation models that are applied to studies of demographic phenomena.”

Our new paper “investigates the political role of uncertainties in demographic projections by employing a multimodel approach in two case studies.” Let’s take a look.

The Song Jian Lesson

In the late 1970s, Song Jian — a Chinese missile scientist trained in Moscow in control systems theory — traveled to Helsinki for an engineering conference and encountered Dutch mathematicians who had adapted missile guidance equations to population dynamics. Song took the framework back to China, applied it to Chinese demographic data, and produced projections showing China’s population exploding past 4 billion by 2080. Those projections drove the one-child policy, adopted in 1980.

Song’s model targeted an “optimal” population of 700 million by 2080 — framed as a missile guidance problem: what fertility trajectory, applied year by year, steers the population to the desired endpoint?

We reproduced Song’s model and ran 256 simulations, varying eight parameters — initial age structure, mortality trends, sex ratio at birth, and optimization settings — across plausible ranges.

Nearly every simulation converged to 700 million. Even though we changed the inputs, the same answer emerged. The optimization constraints left the model almost no freedom to produce anything different, as shown in the figure below.

Stanford economist Paul Pfleiderer terms this a chameleon model, in the context of economic modeling:

“A model becomes a chameleon when it is built on assumptions with dubious connections to the real world but nevertheless has conclusions that are uncritically (or not critically enough) applied to understanding our economy. As will be discussed below, I call these models chameleons because they change colours in order to avoid having their assumptions subjected to appropriate scrutiny. Chameleons are not just mischievous. They can be harmful—especially when used to inform policy and other decision-making—and they devalue the intellectual currency.”

Think RCP8.5 . . . but I digress.

Song’s model was a chameleon model: The conclusion followed not from the inputs but from the model’s design — the sophisticated methodology provided an illusion of scientific rigor.

A more useful decision-support model responds to its inputs — different assumptions produce different results, and those differences carry information about the significance of different assumptions and real-world opportunities and risks.

In China, the consequences were profound.

China’s demographic decline began much faster than Song’s model projected. The one-child policy partially relaxed in 2016. China’s population peaked and began shrinking in 2022.

Current projections place the end-of-century figure at just over 500 million—nowhere near the “guaranteed” 700 million. The model did not reflect reality.

The Garden of Forking Paths

For global projections to 2050 and 2075, we ran four standard demographic models—the Cohort-Component model, the UN World Population Prospects model, the Lee-Carter model, and the Lotka-Volterra model — using UN baseline data and varying key parameters across 1,024 simulations.

The models produce radically different answers, as you can see in the figure below.

The Cohort-Component and UN models land broadly in the range of perturbed UN forecasts. The Lee-Carter model projects well below 4 billion by 2075 in most runs. The Lotka-Volterra model swings to 15–25 billion by 2075 — likely because improving survival rates cause it to overestimate intrinsic growth, a bias the data compound: cohort fertility has not rebounded anywhere it fell below replacement.

The combined output doesn’t form a bell curve around a central estimate. It forms a multi-peaked distribution, four different narratives with vastly different views of the future.

We performed a sensitivity analysis to decompose the variance: Model choice accounts for 96–98% of projected population variance by 2075. Fertility assumptions contribute roughly 2% and assumptions of mortality and migration are just noise.

This result illustrates the ”garden of forking paths”: at a first decision point in demographic research — which model to use? — the answer may already determine most of the answer.

A while back here at THB, I explored the sensitivity of climate projections to population assumptions. That analysis showed that replacing the population assumptions embedded in SSP2-4.5 (“middle of the road”) with more realistic lower trajectories meaningfully reduces projected 2100 temperatures — because cumulative CO₂ emissions scale with population, and the SSP2 population path sits well above defensible demographic projections. The new CMIP7 HIGH projects >14 billion people in 2100 — also implausible

Our new preprint establishes the methodological foundation for that argument. Demographic projections carry structural uncertainty. The SSP population pathways represent one set of modeling choices among many. And the choice that most consistently inflates projections — model selection — is not considered.

The structural uncertainty in demographic modeling skews toward lower populations than the mainstream SSP assumptions embed.

More precisely, the SSP climate scenarios — the ones the IPCC uses to project future temperatures — each utilize a specific population projection. Our analysis shows that when you run multiple credible demographic models, the results spread across a wide range — and the spread does not center on those SSP population figures. The models that track observed fertility trends most closely, and the simple fact that no country where fertility has fallen below replacement has ever seen it recover, both indicate that the plausible range of population futures is trending downward. That matters for climate projections because, all else equal, more people means more emissions — so inflated population assumptions inflate projected temperatures.

That supports taking the under.

Questions to Ask of Any Projection

A projection states a conditional: if these assumptions hold, this outcome follows. A forecast assigns probabilities. A scenario constructs a plausible narrative without asserting likelihood. Conflating these — such as presenting a conditional projection as a forecast— can lead the false confidence illustrated by the Song Jian episode.

Before using any demographic projection as a policy input, ask three questions:

• Does the model respond to its inputs? A model that converges to the same output across varied assumptions reveals a design flaw, and likely not anything empirical about the real world. Sensitivity analysis — systematically varying inputs to observe output change — constitutes basic quality control.

• Is structural uncertainty acknowledged? The dominant uncertainty in demographic modeling doesn’t come from the fertility rate or the mortality table. It comes from model choice — and model choice nearly always goes unreported.

• Does precision reflect real knowledge? A projection showing 9.7 billion in 2050 doesn’t differ meaningfully from one showing 8.9 billion once you account for structural uncertainty. Spurious precision misleads decision-makers who mistake exactitude for accuracy. Climate model projections conditional on single population projections mask a much larger uncertainty, and that uncertainty is greatest on the cooler side.

Our new paper suggests that individual demographic projections tell us more about the models that generated them than about the future.

The preprint and all data and code are openly available.

Before you go: Please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

Comments welcomed, especially on the preprint itself!

Leave a comment

THB exists because of your support. Thanks!

Subscribe now

Last month I was in New York, and sat down with John Stossel who invited me to come talk with him about my experiences in the climate debate over the past 20 years or so. The result is the short video above (just 13 minutes). I am floored by how well done and high quality it is — Already more than 215,000 views. You can also find it on YouTube here.

Please check it out!

While I have you, here is a link to an op-ed I had in The Washington Post last week on RCP8.5 RIP — ungated here.

Subscribe now

Before you go: If you think cancel campaigns are a bad idea please click that “❤️ Like” button to let everyone know. More likes means more readers of THB. Thanks!

Leave a comment

Comments, questions welcomed!

Subscribe now

Two weeks ago I wrote about the most significant development in climate science in decades: the international committee responsible for producing the scenarios adopted by the IPCC formally retired RCP8.5, SSP5-8.5, and SSP3-7.0 — labeling them “implausible.” Last week I documented the near-total silence that greeted the announcement from the English-language mainstream press.

Then President Trump posted about RCP8.5 on social media, calling it “WRONG! WRONG! WRONG!” (just about the only accurate part of his post!).

Within 48 hours, the New York Times, Washington Post, Bloomberg, AP, and Carbon Brief and others published on the scenarios. The main motivation for the coverage centered on Trump, not the new scenarios.

Today, as a follow up to my post noting that there was almost no English language coverage of RCP8.5 RIP, I take a closer look at this first wave of mainstream English language media reports.

In a nutshell: The overwhelming framing of the “climate beat” is that there is really nothing to see here, and to the extent that there is, what we are witnessing reflects the incredible success of climate policy. Right wing media has focused more on the politics, emphasizing the scenario evolution as a “win” for President Trump.

There is also much that is accurate across the coverage, and it is a positive that the news of the new scenarios is getting out. I do not expect the first wave of coverage to be the last.

Let’s take a look at how RCP8.5 RIP has been reported by several major outlets (and readers are invited to add other examples and critiques in the comments).

The New York Times

The NYT story was not framed around about a major development in climate scenario science but instead focused on the president’s social media post. I was interviewed for the story and pretty much everything I told the reporter was not reflected in the story, though I was characterized (some what uncharitably) and quoted (accurately).

The headline and subtitle — shown in the image below — have just about everything incorrect:

• It was not a “tweak” (it was a major revision to the entire scenario envelope);

• Renewable energy was not the reason for RCP8.5 RIP (as shown here at THB);

• It was not a “worst-case” scenario (it was always a baseline/reference scenario);

• Climate scientists were wrong all along (the extreme scenarios were always implausible).

It is true however that “Trump Weighed In.”

Eric Niiler, the NYT journalist, asked me what I think his readers should know. It is a good question. He heard my answer but his readers didn’t.

I told him that the retirement of the extreme scenarios should lead responsible outlets to engage in journalistic accountability, and especially the NYT which has fed its readers a steady diet of overhyped stories on RCP8.5. Not surprisingly, Mr. Niiler has been a participant — Avalanches! and End of the Winter Olympics! — I’m sure everyone on the NYT climate beat has contributed as well to the RCP8.5 coverage.

The NYT will eventually move away from nothing to see here but it might take a while.

The Times (UK)

The Times (UK) was among the first mainstream English-language outlets on the story. That’s the good news. Unfortunately, science writer Adam Vaughan filled the story with errors:

An update to the scientific scenarios used to predict Earth’s future climate reveals a narrowed window of possibilities, and the most apocalyptic worst-case scenario is ruled out thanks to the rapid rise of renewable energy.

Scenarios envisaged by climate scientists in 2010 assumed that the most catastrophic possibility was a rise of about 4.5C above pre-industrial levels by 2100. However, that is no longer considered likely. A new high-emissions scenario projects about 3.5C of warming by 2100.

Not “worst case” and not caused by “renewable energy” and not “4.5C.”

He also gets the numbers wrong.

• According to IPCC AR6 (Table 4.2, global increase over 1850-1900) = 4.8C

• The new CMIP7 MEDIUM = 2.6C

These numbers are apples to apples: SSP5-8.5 was a baseline and CMIP7 MEDIUM is “current policies.” In Vaughan’s favor, he was told by a CMIP7 ScenarioMIP member that the proper comparison to RCP8.5 is CMIP7 HIGH and also misrepresented its value as 3.5C. Both assertions are simply wrong.

To help make things clear, I created the table below to show the proper comparison between the RCPs and the CMIP7 for their baseline/current policy scenarios.

Vaughan’s article also referenced me, describing me as a “climate blogger.” Vaughn, who has come after me in print in the past when he was at The Guardian, might have also noted my side hustle as a researcher whose extensive peer-reviewed research on climate scenarios was favorably cited by the IPCC AR6, foreshadowing RCP8.5 RIP.

Associated Press

The AP story by Seth Borenstein, contains similar errors, which you’ll now recognize:

The Paris climate agreement in 2015 set a goal of limiting warming to 1.5 degrees Celsius (2.7 degrees Fahrenheit) since pre-industrial times, or the mid-1800s, giving rise to the mantra “1.5 to stay alive,” but now scientists say that even their best case scenario still shoots past that signature temperature mark. On the other end, those same new scenarios no longer include the coal-heavy future that would lead to 4.5 degrees Celsius (8.1 degrees Fahrenheit) of warming by 2100, a scary scenario that many scientific studies used in their future projections.

The new proposed worst case scenario has an end-of-the-century warming of about 3.5 degrees Celsius (6.3 degrees Fahrenheit), a full degree (1.8 degrees Fahrenheit) less than the old scenario.

To Borenstein’s credit, he does include and accurately reflect my views:

American Enterprise Institute’s Roger Pielke Jr. said changes to the highest end scenario matter because it was presented as a likely future that could come true if nothing changed. Thousands of scientific studies have been based on that highest warming scenario, called RCP8.5, even though research had already shown it to be improbable.

“It was always presented as where we were headed absent explicit climate policy,” even though it was based on out-of-date and incorrect coal-heavy energy theories, Pielke said in an email.

Interestingly, he quotes Keywan Riahi, lead researcher of the team that developed RCP8.5: “It was never a likely case.” Oh, really?

Perhaps that is what he believes now or even 15 years ago, but it leads directly to the question, why did no one shout out when RCP8.5 was made the only baseline scenario for CMIP5 and IPCC AR5 and everything that followed in policy? When the even more extreme SSP5-8.5 was made a baseline for CMIP6 and IPCC AR6?

Baseline scenarios play a very unique role in research and policy. They are not “worst case” or exploratory research tools.

Other Media

Bloomberg, Carbon Brief, and The Washington Post had similar stories with similar errors, which I won’t detail here.

Common to the coverage by these outlets is a framing that seeks to minimize the changes to the scenarios, attribute those small changes to renewable energy, and to focus on critiquing President Trump — who has nothing to do with scenario development or use, but he does take up a lot of oxygen.

These are the same outlets that have spent years running stories premised on RCP8.5, so it is not surprising that their first wave of coverage is to minimize and deflect.

Right Wing Media

Fox News covered RCP8.5 RIP and also led with the Trump angle. Fox News also repeated many of the same errors, calling RCP8.5 a “worst case scenario” that

“is being phased out after researchers concluded it no longer reflects the most plausible trajectory based on renewable energy growth, emissions trends and climate policies.”

Wrong.

The article’s focus is politics with several quotes from Trump, Hillary Clinton, and Lee Zeldin, the EPA Administrator.

The Daily Caller had a long article that goes into more scientific detail, and while repeating some of the errors common to the outlets above, also quotes extensively from THB, so gets those parts right!

Summary

The first wave of mainstream English-language coverage of RCP8.5 RIP shares a common framing: the changes are minor, renewable energy is the cause, and the real story is Donald Trump. This is all spin.

The outlets that spent years amplifying RCP8.5-based claims have so far chosen to minimize and deflect rather than reckon with that record. It would be unrealistic to expect accountability journalism in the first weeks, but it will come.

Before you go: If you think the mainstream media has some accountability tasks ahead, please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

Comments, questions, discussion, debate — All welcome!

Leave a comment

Subscribe now

Subscribe now

When my son was 12 years old he was 5 feet tall. By the time he was 16 he was 6 feet — growing a whole foot taller in just 4 years. Having a math degree, I felt like I should create a quantitative scenario of his future growth to ensure that everything was OK.

I quickly became alarmed. Based on my calculations, I estimated that he would be 9 feet tall by age 28, growing at a rate of a foot every four years!

I spoke to a doctor, who said to make sure he got a balanced diet, good rest, and regular check-ups. We did all of these things and they worked!

My son topped out at 6’ 2” and, thanks to my alarming growth scenario and quick intervention, the worst case was avoided — he did not grow to 9 feet tall.

What is wrong with this story?

My original 9-foot-tall scenario was never plausible. The dynamics it describes are not how things work. So even though we took good care of my son, that is not the reason he did not grow to 9 feet tall. The scenario was implausible from the start, and my self-described heroic role in averting that scenario is an incorrect reading of that history.

If you understand this little analogy, then you understand current responses from some climate scientists to the retirement of RCP8.5. No, RCP8.5 did not become implausible because of climate policy. Today, I explain why.

In the past several weeks prominent climate researchers have defended RCP8.5 as a scenario that a decade ago plausibly described where the world was headed, but thanks to their warnings, the world’s policy makers responded with implementation of climate policies that have now made RCP8.5 implausible. The implication of these claims is that the world was once headed for ~4.8C temperature increase by 2100 and now it is ~2.7C — a huge decrease.

For example:

• Detlef van Vuuren,1 lead author of the ScenarioMIP paper released last month, explained to The Australian that RCP8.5 had,

“become implausible, based on trends in the costs of renewables, the emergence of climate policy and recent emission trends.”

• Zeke Hausfather, a climate scientist at Stripe and a frequent, friendly intellectual sparring partner of mine, also framed RCP8.5 as once plausible but now implausible due to climate policy successes:

“[I]t is incontrovertible that rapid cost declines, investment in, and deployment of clean energy technologies in the past 15 years have changed the plausible scenarios for fossil fuel use later in this century. These new scenarios reflect this success.”2

• Robert Vautard, Co-Chair of IPCC Working Group I for AR7, also framed the retirement of RCP8.5 as the result of successful climate policies:

“Previous “high scenarios” started in 2015 and assumed no climate policies, but there ARE now many climate policies in many countries, developed in particular with the Paris Agreement signed in 2016 (sic), and before. . . it shows that climate mitigation policies do consistently reduce global warming.”

Each of these framings rests on a common logic: RCP8.5 once described a plausible trajectory; subsequent policy progress and technology cost trends moved the world away from it; therefore the scenario became implausible.3

This story, were it true, would be incredibly convenient for the climate science community. Rather than introducing a flawed scenario to the world that dominated climate science and policy for more than a decade — and then stubbornly defending it — this retelling characterizes the climate science community as near-infallible and heroic.

This story is not true. RCP8.5, and other extreme scenarios, were never plausible.

Scenario plausibility is determined by what theory and evidence supports at the time a scenario is created, not simply by whether the world eventually moved toward or away from the projections that emerge from that scenario.

That means that a scenario that deviates from how the world actually evolved was not necessarily retroactively implausible at the time it was created. A scenario built on assumptions inconsistent with available theory and evidence is implausible at construction, regardless of whether subsequent events confirm or contradict its projections.

The Intergovernmental Panel on Climate Change (IPCC) Special Report on Emissions Scenarios (SRES) — published in 2000 — defined a scenario as:

“a coherent, internally consistent, and plausible description of a possible future state of the world.”

Scenarios are explicitly “not predictions.” But they must be consistent with theory and evidence.

Any projection built on scenario assumptions that contradict available theory and evidence is invalid from the start, regardless of what happens next. Further, scenarios are not predictions, and a family of scenarios does not describe a probability distribution of expected futures. Much wisdom on scenarios has been lost since IPCC SRES in 2000.

Below, I discuss three assumptions of RCP8.5 that made it implausible from the start (ignoring other implausible assumptions, like its incredible population growth rates):

• Reliance on a flawed theory for the dramatic expansion of coal energy;

• A corresponding rapid increase in coal-to-liquids, displacing petroleum;

• A necessary slowdown in technological improvements in solar energy technology.

Let’s take a look at each.

First, RCP8.5 required burning coal at an implausible rate. Coal is the most carbon intensive fossil fuel, and huge amounts needed to be burned to reach the high forcing level that was assigned to the most extreme RCP scenario.

Recall that under the RCPs radiative forcing levels for 2100 were chosen first, and only later were integrated assessment modelers assigned the task of figuring out how those levels could be reached. RCP8.5 required implausible socio-economic assumptions to meet its preassigned design criteria. Unfortunately, the scenario develoipment community yet to learn that socio-economics should come first.

The figure above, from Ritchie and Dowlatabadi 2017 (RD17), shows that RCP8.5 assumed ~8x increase in primary energy generation from coal. In fact, the entire envelope of assumed future coal consumption across the RCP and SSP scenarios shows an increase in coal — based on a single theory, characterized by RD17 as a “return to coal.”

RD17 explain that the MESSAGE integrated assessment model that generated RCP8.5 applied no constraint from geological reality, and simply assumed that the massive amounts of coal it required would be available. The scenario extrapolated 2000s Chinese coal growth rates and assumed the physical resource base would accommodate whatever the consumption assumption required irrepective of real-world constraints.

The figure below shows that coal consumption as a proportion of the global energy mix started declining around 2013, while RCP8.5 had it steadily increasing. The reason for the divergence between reality and RCP8.5 was not climate policy, but rather a false assumption baked into the scenario from the start.

RCP8.5 — and indeed all of the RCP and SSP scenarios — had a single point of failure in its assumption of a return-to-coal. This assumption alone settles the question of plausibility. A scenario requiring five times proven coal reserves is not plausible by any standard.4

There is another important conclusion to reach here: The entire set of RCP and SSP scenarios are contaminated by the “return to coal” theory. That means that any use of these scenarios as somehow representing a probablistic forecast of the future is simply flawed. However, the IPCC and others routinely make this mistake.

Second, the assumed dramatic expansion of coal consumption necessarily required additional implausible assumptions, notably the increasing use of coal-to-liquids (CTL) to replace petroleum and a slowdown of technological improvements in wind and solar technologies.

Here as well, the assumed increase in CTL was not averted due to climate policies. It represented a flawed assumption of RCP8.5 that was necessary to meet its coal growth assumption. The assumed increase in CTL was implausible at the time RCP8.5 was created.

Third, the assumed coal expansion required still more implausible assumptions. For example, RD17 explain that the assumption of a dramatic expansion of coal energy necessarily required other energy technologies to stagnate in their technological and economic improvements:

A coal-dominant energy system in RCP8.5 results from coal investment costs that continually decline, while the learning curve for solar, wind and nuclear power remain static.

For example, the figure below shows that the RCPs each required that solar PV experience a slowdown in improvements in module costs, as compared to the rate observed historically. In the figure, the RCPs are initiated in 2011, the year of their release.

Those who attribute RCP8.5’s implausibility to “trends in the costs of renewables” are correct that these costs diverged from what the scenario projected. But they have causation backwards: those cost declines were the continuation of a fifty-year trend that suggested a strong baseline.

The scenario’s failure to capture the continuing trend in declining solar costs was simply the flip side of the implausible coal expansioon assumption. Solar needed to make way for evermore coal.

In Burgess, Ritchie, Shapland, and Pielke 2021 we examined all AR5 baseline scenarios against actual CO₂ emissions across 2005 to 2045, combining observations and near-term energy system model projections. We documented the differences between the RCPs and SSPs that are discussed in this post, as you can see in the figure above from that paper.

In that paper we concluded:

Recent (post-2005) trends and energy outlook projections (to 2040) of global CO₂ emissions are substantially lower than projected by baseline scenarios used in the IPCC’s Fifth (AR5) and Sixth (AR6) Assessment Reports, and are well off-track from widely-cited high-emission marker scenarios such as RCP8.5. We show that this divergence owes largely to per-capita GDP and carbon intensity growth slower than projected in baseline scenarios. The gap between observed and projected carbon intensity is very likely to continue to increase throughout the 21st century due to the implausible assumptions high-emission scenarios make about future fossil-fuel expansion (Ritchie and Dowlatabadi 2017).

The fact that RCP8.5 was implausible has been well established in the scientific literature since 2017. Further evidence has refined that view in the years since.

What happened after the RCPs were released in 2011 — Paris, the renewables revolution, expansion of US shale — is the unfolding history of the world continuing not to be plausibly characterized by RCP8.5. The scenario did not become implausible. The evidence that it was implausible simply became undeniable as the real world and the RCP8.5 world continued to diverge.

It is important that the community understands how the RCP mess came about and take steps to ensure that it does not happen again.

I fully understand both the need to save face and to portray climate policies as being more effective than they actually have been. However, if we do not accurately understand the reasons for the RCP8.5 debacle then we risk repeating those mistakes.

Before you go: Please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

Comments, questions, discussion, debate — All welcome!

Leave a comment

Subscribe now

Read more:

• Today—>Hausfather, Peters, Forster: On the death of RCP8.5

• Today—>Burgess: Thank you for your attention to this (RCP8.5) matter

• SRES: Nakićenović, N. & Swart, R., eds. (2000). Special Report on Emissions Scenarios. IPCC/Cambridge University Press.

• Van Vuuren, D.P. et al. (2026). ScenarioMIP-CMIP7. Geoscientific Model Development, 19(7), 2627–2656.

• Hausfather, Z. & Peters, G. (2020). Emissions — the ‘business as usual’ story is misleading. Nature, 577, 618–620.

• Hausfather, Z. (2025). An assessment of current policy scenarios and the reduced plausibility of high-emissions pathways. Progress in Physical Geography, 49(1).

• Ritchie, J. & Dowlatabadi, H. (2017). Why do climate change scenarios return to coal? Energy, 140, 1276–1291.

• Burgess, M.G., Ritchie, J., Shapland, J. & Pielke, R. (2021). IPCC baseline scenarios have over-projected CO₂ emissions and economic growth. Environmental Research Letters, 16(1), 014016.

• Way, R., Lafond, F., Lillo, F. & Farmer, D. (2022). Empirically grounded technology forecasts and the energy transition. Joule, 6(9), 2057–2082.

• Riahi, K. et al. (2011). RCP 8.5 — A scenario of comparatively high greenhouse gas emissions. Climatic Change, 109, 33–57.

• Global Carbon Project (Friedlingstein et al., annual). Global Carbon Budget.

• Energy Institute (2024). Statistical Review of World Energy.

Subscribe now

Suppose you ask a meteorologist for a weather forecast to help with planning a picnic for tomorrow. She gives you a detailed, technically sophisticated weather forecast for Mars.

Every equation underlying the forecast is correct. The modeling is state of the art. But the forecast is irrelevant to your picnic planning, because the forecast describes the weather on the wrong planet.

For well over a decade, a large portion of climate research and the use of that research has had a real-world version of this problem. The scenarios driving climate projections — the foundational assumptions about our collective future — described a world so removed from plausible reality that the projections built on them tell us more about a hypothetical future than about the one we are actually navigating.

The news that the most extreme climate scenarios have now been officially put out to pasture has now begun to spread far and wide. The scenarios — specifically, RCP8.5, SSP5-8.5, and SSP3-7.0. — were quietly retired last month by the international committee responsible for developing a new basket of official scenarios. It cannot be overstated how significant this change is — the now-obsolete extreme scenarios underpin the work of the Intergovernmental Panel on Climate Change (IPCC), tens of thousands of research papers, government policy and regulation around the world, financial standards for the world’s banks, along with much of the media coverage of climate change, from which most people learn about climate science and policy.

Longtime readers of THB, and the outlets where I published before, will know that my colleagues and I have called for the retirement of the extreme scenarios for almost a decade. Now it has happened and the fallout inevitably will be significant.

Back in 2019, building on the foundational work of my collaborator Justin Ritchie, he and I set about trying to document and make sense of how it came to be that the climate research community became locked in on scenarios that were fundamentally flawed and implausible — distorting our view of the climate future.

The result was a magnum opus paper, coming in at more than 21,000 words, published in 2021:

Pielke Jr, R., & Ritchie, J. (2021). Distorting the view of our climate future: The misuse and abuse of climate pathways and scenarios. Energy Research & Social Science, 72, 101890.

Our paper explains the many mistakes made in the development of climate scenarios that have come to have profound global impact, far beyond the science of earth system modeling.

As we quote one participant in the scenario development process:

“You simply do not realize that the RCPs can start a life of their own.”

I have received dozens of requests to help people better understand the significance of the extreme scenarios, why their retirement is such a big deal, and how we got into this mess in the first place. Today, I summarize Pielke and Ritchie 2021 with the remainder of this post, and for THB paid subscribers at the bottom I offer a PDF of the full text of the paper.

It is a remarkable story.

Scenarios Are the Foundation of Everything

Long-term projections of the climate future depend on much more than just math and physics— they are built upon a foundation of internally consistent stories about plausible futures, called scenarios. Scenarios project answers to question such as: How many people will there be? How wealthy will they be and in what occupations? What will power the economy? What technologies will we have for agriculture, industry, transportation and so on?

These scenarios result in projections of greenhouse gas emissions, land use and land cover, aerosols, and many other influences on the climate system that feed into earth system models, which produce projections of variables such as temperature, sea level, drought frequency, and storm intensity. The results of these projections are typically fed into still more models that project climate impacts, economic costs and benefits, and the possible consequences of alternative policy options.

Those projections do not stay in academic journals. Financial regulators in Europe, the United Kingdom, and the United States now require banks and insurers to conduct climate stress tests built directly on this work. Infrastructure engineers consult scenarios when setting design standards for roads, bridges, ports, and water systems. Urban planners use them to decide where to permit development. The insurance and reinsurance industry prices risk from them. Credit rating agencies assess sovereign and corporate debt through futures envisioned via scenarios. And much, much more.

When the scenarios are wrong, everything downstream is wrong too. They propagate into trillion-dollar investment decisions, regulatory frameworks, engineering specifications, and government policy.

Climate Modeling and Decision-Making Need Different Things

Our paper documents a central tension: climate modelers and real-world decision makers need fundamentally different kinds of scenarios, and the scenario development process has consistently prioritized the needs of researchers over the needs of decision makers.

Climate modelers want to employ a wide range of inputs to their earth system models — from very low to very high atmospheric greenhouse gas concentrations — so they can map the climate system’s response across a wide range of inputs. An extreme high-end scenario is very useful for research seeking to detect forced signals against the noise of natural variability, and for comparing model outputs across different research groups. Extreme scenarios generate large, clear changes that are easier to identify and analyze.

Whether that extreme scenario is plausible in the real world is, for the climate modeler’s specific technical purpose, largely beside the point.

Decision makers need something completely different. A city engineer designing a flood barrier, a central bank stress-testing a loan portfolio, an insurer pricing hurricane risk — all of them need scenarios grounded in the real world because that is where their decision making takes place.

They often want to have a credible baseline or “current policy” scenario that offers a defensible account of where the world — as it is today — is likely heading without major new policy interventions. Such a reference scenario should be connected to actual trends in energy technology, economic development, and demographics. For decision makers, an implausible scenario is worse than useless — they produce misleading numbers that could very easily lead to misdirected investment, distorted regulation, and flawed planning.

For most of the history of climate science, researchers managed this tension by developing socioeconomic scenarios first, as illustrated in the figure below. Demographers, technologists, economists and others together constructed internally consistent stories about future human society, those stories drove emissions projections, and those projections then fed into climate models as inputs.

The socioeconomic foundation came first; the physical climate science followed. That sequence mattered because it anchored the entire chain in a recognizable account of the human world — one that could at least be interrogated and debated by the engineers, planners, economists, and regulators who ultimately had to act on the results.

The Plausibility Vacuum

The creation of the Representative Concentration Pathway (RCP) scenarios, developed starting in 2005, broke that sequence. The design intent was a “parallel approach”: rather than waiting for socioeconomic scenarios to be built first, climate modelers would receive radiative forcing pathways immediately — atmospheric greenhouse gas concentrations specified over time — so they could begin their long, computationally expensive model runs without delay. The socioeconomic scenarios would follow later and — hopefully — plausibly lead to the radiative forcing trajectories that had already been adopted without consideration of their plausibility.

This approach created what Ritchie and I call a “plausibility vacuum.” Once radiative forcing was severed from its socioeconomic foundation, there was no longer any mechanism to ask whether a given forcing pathway was consistent with a plausible description of human society.

The RCPs floated free. Researchers using them to project impacts on agriculture, public health, ecosystems, and infrastructure could no longer be certain that the human world embedded in their scenarios bore any coherent relationship to the physical climate outcomes they were projecting. For research focused on better understanding the physical sciences of climate and climate change, the plausibility vacuum didn’t matter. But for everyone else, it did.

The parallel approach that created the plausibility vacuum was originally sold as a temporary measure. In practice it has been permanent — Consider that since the start of 2025 about 7,500 research articles have been published with RCP8.5.

When socioeconomic analysis finally caught up years later it revealed that the most commonly used scenario — the one whose creators labeled “business as usual” — required levels of coal consumption and population growth that were implausible. In fact, the integrated assessment models that produced the other three RCPs — RCP2.6, RCP4.5, and RCP6.0 — could not even reach a radiative forcing of 8.5 watts per square meter, under any set of input assumptions.

The selection of RCP8.5 as the highest priority scenario for climate research, and the only one designated as a reference or baseline scenario, was a fateful choice. Remarkably, climate scenario development in 2026 still takes place in a plausibility vacuum.

The RCPs Were Never Comparable with Each Other

The problems run deeper still, because the original four RCPs were not derived from a common framework. Each came from a different integrated assessment model — IMAGE, MiniCAM, AIM, and MESSAGE — which were developed by different research groups working from different assumptions about population, economics, technology, and land use. Each model had its own internal baseline, against which its own policy interventions produced lower forcing outcomes.

The four scenarios were never apples-to-apples. They were four different fruits from four different trees. Yet, over more than a decade and in tens of thousands of papers, RCP8.5 was treated as where the world was headed and the other three scenarios — but especially RCP4.5 and 2.6 — as a world with climate policy interventions. The 2018 U.S. National Climate Assessment treated RCP8.5 as a reference and RCP4.5 as policy success, both assumptions were wrong.

The RCP designers warned the community about this explicitly when the scenarios were being developed — the scenarios “cannot be treated as a set with consistent internal logic” and the high scenario “cannot be used as a no-climate-policy reference scenario for the other RCPs.” That warning has been comprehensively ignored.

The Scenario That Took Over

One of the four RCPs — RCP8.5, the highest — came to dominate the literature to a degree that is impossible to overstate. RCP8.5 accounted for more than half of all RCP references in the 2018 U.S. Fourth National Climate Assessment, nearly 60 percent in the IPCC’s Special Report on the Ocean and Cryosphere, and about a third of all RCP references in the IPCC Fifth Assessment Report.

By early 2020, researchers were publishing studies invoking RCP8.5 at a rate of roughly 20 per day. So far in 2026, studies using RCP8.5 (or its even more extreme successor, SSP5-8.5) are being published at a rate of ~30 new studies per day.

The dominance of RCP8.5 happened for reasons that are explainable and understandable, even if deeply pathological. Here are some of those factors, and certainly not an exhaustive accounting:

• When the four RCPs were published, only RCP8.5 was structured as a baseline — the other three were all constructed as policy intervention scenarios. Researchers who needed a no-policy reference had one option available.

• The IPCC, which assesses published literature, consequently emphasized RCP8.5 in its reports.

• Media coverage amplified the click-friendly, alarming projections from the RCP8.5 studies.

• Researchers may or may not have cared about plausibility of the scenarios that underlay their research, but for those whose careers depend on publication and visibility it was no doubt a feature not a flaw that RCP8.5 generated the most striking results attractive to journal editors, climate beat reporters, and university press offices.

Justin and I explain that no one need invoke bad faith or a conspiracy:

The bottom line is that scenario misuse involving the RCPs resulted from myriad factors coming together and reinforcing each other. They range from the ridiculously simple – the common naming scheme for the RCPs, to the incredibly complicated – the collapsing of complexity involved with the notion of baseline scenarios in methodologies of scenario planning, abuse of the scenario probability vacuum, to institutional dynamics – the IPCC assuming the role of orchestrating the very literature that its main function was simply to assess. As such the objective of understanding scenario misuse is not to apportion or assign blame, but to understand how such a pervasive and consequential failure of scientific integrity came to be on such an important topic, how it can be corrected and how it can be avoided in the future.

The Successor Scenarios Did Not Fix the Problem

The Shared Socioeconomic Pathways (SSPs), published in 2017 were supposed to restore the socioeconomic foundation that the RCP process had severed. They did so only partially and too late. The SSP development process actually confirmed that a forcing level of 8.5 W/m² “can only emerge under a relatively narrow range of circumstances” — language that should have triggered a fundamental reorientation of the research agenda. Instead, the desire for continuity with a decade of prior modeling work meant that SSP5-8.5 was designated the highest-priority scenario for the climate model experiments informing the IPCC Sixth Assessment.

The modeling community’s need for continuity — for results that can be compared to earlier model runs — trumped the policy community’s need for scenarios grounded in the real world. That is the dynamic the scenario development process has consistently reproduced, across multiple generations of scenarios.

The SSPs also introduced a new pathology. Researchers began mixing elements from incompatible scenarios — combining the grim, impoverished-world narrative of SSP3 with the extreme forcing level of RCP8.5 to construct a “chimera” scenario, SSP3-8.5, that the SSP developers themselves had flagged as implausible. Dozens of published studies now use this combination to explore worst-case impacts, generating projections built on a future world that no serious socioeconomic analysis supports.

Our paper goes into some detail on the SSPs. Those interested can see more there.

What Needs to Change

Course correction will be difficult. Enormous institutional momentum is well entrenched — thousands of published papers, active grants, ongoing IPCC cycles, regulatory frameworks already built on RCP8.5-derived projections. But the direction of change is clear, the only question is how long it’ll take to get back on course.

Over the week or so since I announced to the world that the extreme scenarios have been retired, my social media feed have been filled with many, including experts, with some version of “nothing to see here.” In a future post I’ll chonicle and correct the many false claims that are being spun about the retirement of the extreme scenarios.

There is much that needs to be done to correct course in climate science and policy. For the research community focused on meeting the needs of decision makers, near the top of the list: scenario development needs to be more frequent, more anchored to near-term policy-relevant time horizons, and more accurate about what real-world trends actually imply — something closer to how the International Energy Agency updates its scenarios annually in light of current conditions.

Most fundamentally, the needs of decision makers must be given equal weight to the needs of climate modelers in the scenario development process. Those two audiences require different things, and consistently privileging one over the other has produced a decade of science that is technically sophisticated and often policy-irrelevant, if not flat-out misleading.

Even better, exploratory climate research should be spun off from that focused on informing decision makers. We simply cannot kill two birds with one stone.

Why This Matters

Climate change is real. The risks are serious. The case for strong policy action does not depend on whether RCP8.5 is a plausible baseline — All of the arguments I made back in 2010 in The Climate Fix survive the retirement of extreme scenarios.

Ultimately, successful climate policies necessarily require broad public confidence in the integrity of research and demonstration that science is self-correcting. How the community responds now will go a long way to determining whether trust is deserved.

Our 2021 paper is a careful account of how the climate science community ended up in this mess — and what it would take to get out. We are not there yet.

Leave a comment

Before you go: If you think more people should know the real story behind RCP8.5 RIP, please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

If you find the work of THB valuable, please consider a paid subscription.

Read more

The annual subscription price for new THB subscribers increases to $100 on June 1.

Subscribe before then and lock in the current rate — permanently.

Subscribe now

Last week here at THB, I published a post announcing the most significant development in climate science in decades. It is truly huge news.

The international committee responsible for official IPCC scenarios had declared the high-end scenarios — RCP8.5, SSP5-8.5, and SSP3-7.0 — to be implausible. These scenarios have dominated climate research, headlines, and policy for the better part of two decades.

Today I review who in the “mainstream” media has covered this major story and who has so far ignored it.

The Dutch media is out in front

The most substantive mainstream coverage came from the Netherlands — perhaps fittingly, since Detlef van Vuuren, lead author of the ScenarioMIP paper that announced the new scenarios and a fixture across generations of climate scenarios, works at Utrecht University and the PBL Netherlands Environmental Assessment Agency.

• De Volkskrant, one of the country’s largest outlets, ran the story on its front page on May 4 under the headline: UN Climate Panel Drops Doomsday Scenario. The story notes that a few years ago De Volkskrant did a self-audit of its own climate coverage and identified 54 articles it had published on RCP8.5 studies.

• Science journalist Maarten Keulemans, who wrote that story, posted on X: “This is so huge. Mind-blowing. Crazy. The IPCC admits what’s been circulating for a while: the highest doomsday scenario, 8.5, no longer matches reality. ALMOST EVERYTHING YOU READ ABOUT THE CLIMATE FUTURE IS WRONG.”

• Van Vuuren was quoted in De Volkskrant and his comments were notable. The consequences of 3.5°C warming are “vervelend genoeg,” bad enough already.

• Van Vuuren characterized the new high-end warming in 2100 as 3.5C, which is considerably higher than the ~3C that I estimated from the available data that the ScenarioMIP posted online and using the same climate emulator. Interestingly, Van Vuuren’s framing — centered on the high scenario, rather than the medium “current policy” scenario — misuses the new high end scenario in a manner that the paper he led said to avoid: by using it as a projective reference scenario, rather than an exploratory “what if?” exercise. I am sure we will be seeing more of this sort of misuse of HIGH. Everyone loves the most extreme scenario available.

• Van Vuuren attributes the need to retire the upper end scenarios to changes in the real world rather than basic flaws in the scenarios. As THB readers well know, this is just wrong. The high end scenarios were always off target, because they were based on flawed assumptions of a world that was going to dramatically expand coal use. Van Vuuren explained to De Volkskrant: “The world has fortunately developed. Renewable energy has become cheaper quickly. And, even if it is still too little, there is climate policy.”

• Credit to Van Vuuren for acknowledging that the elimination of the extreme scenarios will be very disruptive: “That all sorts of forecasts will now have to be revised is difficult. But it obviously cannot be the case that we fail to adjust our scenarios for that reason.” He also mentions that the low scenarios have increased.

• He added: “And if we do too little about greenhouse gas emissions, you will automatically still end up at higher values. It just happens later, after 2100.” I guess that leaves the door open for a 2150 or 2200 return to RCP8.5.

• NPO Radio 1 broadcast a segment featuring Marcel Crok, of CLINTEL (which has graciously translated several THB posts into various languages), who had argued since 2018 that the extreme scenarios were untenable. The segment’s title is: “It became a favorite scenario of scientists.”

German-language coverage

• The Berliner Zeitung ran a balanced piece, arguing that “extreme climate scenarios played too large a role in public debate for too long” — while cautioning that the retirement of RCP8.5 does not mean climate research was fundamentally wrong.

• The Swiss outlet watson.ch covered it under the headline “World Climate Council Drops the Worst Scenario RCP8.5” — repeating van Vuuren’s “vervelend genoeg” as “schon schlimm genug” (bad enough already) and noting Zeke Hausfather’s warning that warming could still exceed 3°C.

• Die Welt also covered the story. The article’s subtitle is hard hitting: “A lobby made RCP8.5 famous: The most sensationalist of all climate scenarios has determined scientific studies, media and politics — yet it is unrealistic. Now it is actually being phased out.”

English-language climate silence

Remarkably, there has not been a peep from the major U.S. or international news outlets that publish in English.

• The New York Times has been perhaps the most prominent English-language home for promoting news stories based on studies that rely on RCP8.5. They’ve said nothing about its retirement.

• The BBC ran a More or Less episode on the RCP8.5 problem in December 2025 — Why did the climate change model get it wrong?. Kudos to them, but they have published nothing on the cancellation of the extreme scenarios.

• Carbon Brief — the specialist outlet that has covered RCP8.5 more than perhaps any other English-language publication — has been silent.

• Neither Science nor Nature has covered the retirement of the extreme scenarios. So far in 2026, more than 2,600 studies have been published using the high-end scenarios and tens of thousands before that — so this is a really big story within science.

• The Guardian, which ran many dozens (hundreds?) of stories premised on RCP8.5 projections over fifteen years, has run nothing.

Charitably, we might attribute the lack of English-language coverage to lag time or the story’s technical nature. Of course, those issues were not a problem for the Dutch and German media.

Less charitably: the outlets most invested in their longstanding promotion of RCP8.5 have the most to lose from a clear-eyed accounting of what its retirement means for science, policy, and their own coverage.

The fallout from the retirement of extreme climate scenarios has not yet begun. Watch this space.

Before you go: If you think more people should know about RCP8.5 RIP, please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

Comments, questions, conversation, critique, and debate — All invited!

Leave a comment

If you find the work of THB valuable, please consider a paid subscription.

Subscribe now

Subscribe now

The list of scientific institutions that have taken overt partisan positions — or stances that create a strong impression of partisanship — is notable:

• In the run-up to the 2020 presidential election, Nature, Scientific American, the Lancet, the New England Journal of Medicine, and Science published editorials criticizing Donald Trump and urging voters to replace him, marking the first presidential endorsements in the histories of several of these publications.

• Eighty-one American Nobel laureates endorsed Joe Biden.

• Four years later, Nature and Scientific American endorsed Kamala Harris.

The 2017 “March for Science” brought more than a million participants to rallies worldwide. More than 100 scientific organizations endorsed the march — among them the American Association for the Advancement of Science, the American Geophysical Union, the American Psychological Association, the American Anthropological Association, and the Union of Concerned Scientists.

Organizers insisted the march was political but not partisan — after all, what could be partisan about marching for science?

However, the reality was that the “March for Science” was less about science and more about marching against President Donald Trump. A survey of participants in the march at three locations published in Science Communication, found that among the 110 participating scientists surveyed, exactly one — <1% — self-identified as Republican, while ~72% identified as Democrats, as you can see in the figure below.

At best, the march was a form of stealth partisan advocacy, endorsed by more than 100 scientific societies.

A study of the effects of the march on public opinion — published in PS: Political Science & Politics — found that the march reinforced polarization:

“Liberals’ attitudes toward scientists became more positive whereas conservatives’ attitudes became more negative.”

The authors concluded:

“this research calls attention to the possibility that the political actions of scientists can shape public opinion about them.”

New empirical research finds that partisan political advocacy by authoritative institutions comes at a heavy price paid in diminished overall public confidence — not just among conservatives but also independents — as well as increased polarization.

None of this has happened in a vacuum. Trust in scientists, which spiked during the early COVID-19 pandemic, has since eroded sharply and unevenly. The share of Americans expressing a great deal of confidence in scientists fell from 39% in 2020 to 23% by 2023, while those expressing not too much or no confidence rose from 12% to 27% over the same period.

A modest recovery followed — Pew’s 2024 survey found overall confidence at 76% — but the deeper pattern is accelerating partisan divergence: 88% of Democrats express confidence in scientists versus 66% of Republicans, and Republicans remain far more likely to oppose scientists’ active engagement in policymaking. Gallup’s 2025 confidence survey placed science at 61% overall — higher than other institutions, but historically low.

Starting in 2005 at the University of Colorado Boulder, I embarked on a project to host public events and interviews with all of the (then) living current and former science advisors to the U.S. president — Conversations with Presidential Science Advisors. We hosted science advisors who served presidents from John F. Kennedy through Barack Obama and every one in between.

In 2006 I interviewed Frank Press who had served as President Jimmy Carter’s science advisor in the 1970s. After that, in the early 1980s he became the president of the U.S. National Academy of Sciences. I asked him why, during his tenure, the National Academy of Sciences never undertook a study of the scientific and technical aspects of President Ronald Reagan’s proposed Strategic Defense Initiative — SDI, popularly known as “Star Wars”.

His answer has stuck with me for 20 years.

Press explained that a sign-on petition opposing SDI had circulated among the scientific community, drawing signatures from roughly sixty percent of Academy members. He said that he believed that the petition had compromised the Academy’s ability to lend an independent voice to the debate as any NAS report — even if put together by a balanced expert committee — would face dismissal in the political process because the institution had seemingly already signaled its position on the political debate.

A potentially valuable independent assessment of the SDI proposal never materialized. The cost of the sign-on petition was paid in the inability to secure legitimate, authoritative expert advice. We all paid that price.

Empirical support for Press’ intuition can be found in a 2023 preregistered large-sample experiment published in Nature Human Behaviour. That study explored the public opinion consequences of Nature‘s 2020 endorsement of Joe Biden. It found that the endorsement caused large reductions in stated trust in Nature among Trump supporters, lowered their demand for COVID-related information from the journal — including substantially reduced requests for vaccine efficacy articles — and reduced their trust in scientists generally. Effects on Biden supporters’ trust were positive, small, and mostly statistically insignificant.

The people who were going to vote for Biden anyway liked the endorsement. Those who weren’t did not. More polarization was the result. It turns out that a science journal in London really doesn’t move voters in the U.S. with an endorsement. But the endorsement nonetheless had effects.

Even with the results of this study in hand, in 2024 Nature responded by doubling down, issuing an endorsement of Kamala Harris. I discussed Zhang’s findings in two posts at THB — November 2023 and April 2024 — and noted that Nature received clear experimental evidence of self-inflicted harm on the community that it serves and then — chose to continue inflicting that harm.

A new paper, just published in Group Processes & Intergroup Relations (ungated preprint at PsyArXiv), significantly extends and reinforces the earlier conclusions. Three pre-registered experiments, 6,281 participants, spanning both the 2020 and 2024 elections tested five journals in 2020 (the same ones that issued endorsements) and three in 2024. Participants read either a neutral description of a journal’s history and mission or the same description alongside excerpts from its actual presidential candidate endorsement text.

Participants who read the endorsements reported lower trust in science as an institution, stronger belief that science lacks impartiality, higher general distrust, greater agreement that scientists’ personal values shape their findings, and reduced confidence that science benefits society. The 2020 studies found endorsements also increased conspiracy beliefs about COVID-19 and the scientific enterprise itself.

The figure below converts findings from the paper’s three studies to a 0–100 scale (from its original 1–9 scale) and shows the change in trust resulting from the endorsements, by political orientation:

The study found that decreasing trust occurred not just among self-described conservatives, but also moderates. Self-described liberals showed small responses, and without statistical significance.

To be clear — Nothing in this evidence argues against scientists speaking their minds based on both their expertise and their citizenship. As THB readers know well, I strongly believe in academic freedom and that academic freedom includes the freedom to speak politically. I do it all the time.

A journal holds a very different position than an individual expert: It is an institution of expertise. Its authority derives from its claim to serve science — and through science, to serve the public, all of the public whether far left, far right, or anywhere in between. When an authoritative scientific institution — whether a journal, university, or national academy — uses its authority in service of a subset of the public and opposed to another subset, there can be no surprise that those being opposed (or perceiving being opposed) respond by questioning the legitimacy and authority of that institution.

Of course, the current dynamics within the scientific community can make leadership difficult for those who would eschew overt partisanship. As the scientific community has become more left-leaning overall and prominent voices demand even more partisanship, eschewing partisanship may not be popular among many in the community. For some, partisanship by scientific institutions is a feature and not a flaw. That makes scientific leadership that navigates competing pressures more important than ever.

Recent research on 98,000 academics on X/Twitter found that non-partisan scientists draw the most public credibility, and more intense political expression is associated with less perceived credibility. Yet, even so, many scientists recognize this trade-oof and nearly half post political content anyway, credibility be damned.

The incentives for scientists to express themselves politically in the attention economy diverge sharply from the collective interest of scientific institutions.

In a hard-hitting essay in Skeptic, “The Siren Song of Influence,” Cory Clark and Bo Winegard argue that the attention economy has warped incentives across science more broadly:

[I]n the age of social media, a growing number of scientists appear less interested in the patient work of discovery and more interested in being political influencers. They post partisan hot takes, endorse candidates, and signal ideological commitments, often leveraging their authority as scientists to advance personal or political aims. This doesn’t just alienate half the country (usually the right-leaning half); it undermines the hard-won and always precarious reputation of science as a dispassionate arbiter of truth. The more we scientists insert ourselves into political battles, the more we risk losing the very authority we seek to protect.

The calculus involved with an individual scientist deciding to try their lot as a political influencer is markedly different from that calculus for an institution of science. Somewhere along the way, leaders of scientific institutions seem to have missed the significance of this distinction.

In a 2004 paper in Environmental Science & Policy, I similarly argued that scientists using institutional authority to advance partisan agendas rather than illuminate policy choices trade away the credibility that gives scientific expertise its public value.

A few years later, in The Honest Broker I described the choices scientists and the institutions they inhabit face in how they relate to broader society. Each of the four roles it describes — pure scientist, science arbiter, issue advocate, honest broker — entails real tradeoffs. Individual scientists who choose to become advocates make a legitimate choice, with costs and benefits. The same goes for each of the other roles.

Institutions face choices as well, but the costs and benefits will be different and far more significant than those faced by individuals.

Research now helps us to better understand those costs and benefits. Institutional partisan advocacy — overt or stealth — comes with a price paid in the erosion of public trust and increased polarization over science.

The question that we should ask leaders of institutions of science — who are supposed to serve everyone — Are the costs of partisan advocacy worth it?

Before you go: If you think institutions of science should serve everyone, please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

Comments, questions, conversation, critique, and debate — All invited!

Leave a comment

For further reading

• Syropoulos et al. (2026). Presidential Candidate Endorsements by Scientific Journals Decrease Trust in Science Especially for Moderate and Conservative Americans. Group Processes & Intergroup Relations. Preprint: https://osf.io/preprints/psyarxiv/qn2ha_v2

• Zhang, F.J. (2023). Political endorsement by Nature and trust in scientific expertise during COVID-19. Nature Human Behaviour, 7(5), 696–706.

• Ross, A.D., Struminger, R., Winking, J., & Wedemeyer-Strombel, K.R. (2018). Science as a public good: Findings from a survey of March for Science participants. Science Communication, 40(2), 228–245.

• Motta, M. (2018). The polarizing effect of the March for Science on attitudes toward scientists. PS: Political Science & Politics, 51(4), 782–788.

• Pew Research Center (2024). Public trust in scientists and views on their role in policymaking.

• Alabrese, E., Capozza, F., & Garg, P. (2024). Politicized scientists: Credibility cost of political expression on Twitter. Working paper.

• Clark, C. & Winegard, B. (2025). The siren song of influence. Skeptic.

• Pielke, R.A. (2004). When scientists politicize science. Environmental Science & Policy, 7(5), 405–417.

• Pielke, R.A. (2007). The Honest Broker. Cambridge University Press.

Subscribe now

Subscribe now

Note: This post documents relationships between science advisors, their funders, NGOs, finance, and governments. I make no allegations of wrongdoing or anything close. What I document is a scientific ecosystem that has developed into having profound global economic and political significance, but institutional accountability, transparency, and governance have not developed in parallel — leaving a sizeable good-governance gap. The governance of climate research requires a rethink.

The Coupled Model Intercomparison Project (CMIP), now in its seventh phase, is the international scientific group under the World Climate Research Program1 that oversees official projections of climate futures. CMIP scenarios drive the temperature and emissions projections that anchor every Intergovernmental Panel on Climate Change (IPCC) assessment and much more:

• The projections inform national climate-impact assessments in the United States, the United Kingdom, Germany, Canada, Australia, Japan, the Netherlands, and many more.

• They drive sea-level-rise estimates that local governments use to plan coastal infrastructure.

• They also shape projections of future climate damages, such as in estimates of the “social cost of carbon,” which has played a key role in U.S. federal regulatory analysis.

• The projections also inform the Network for Greening the Financial System (NGFS) scenarios that more than 140 central banks and supervisors are required to use to stress-test commercial bank capital requirements.

• They form the climate diagnostics that the World Bank’s Climate Change Knowledge Portal provides for over 100 client countries, informing national and local decision making related to climate risk.

It is no exaggeration that the CMIP climate projections influence trillions of dollars in investment and regulation. They are, in functional terms, among the most consequential 21st century scientific products designed to inform policymaking, economics, and regulation.2 They are not just about science, but about science advice to policymakers in government, business, and civil society.

Last week I revealed that CMIP had released illustrative versions of its new scenario set and that it had retired — finally — the out-of-date and implausibly extreme RCP8.5 scenario that has dominated climate research and policy for over a decade.

You might wonder: Who produces this critical scientific guidance for policy? Who participates? Who does not? Who decides who gets to participate? To whom are they accountable? Who supports their work? How is quality control ensured? Who decides what values are prioritized — who wins, who loses in the scenarios?

Remarkably, answers to questions such as these are not at all easy to find, if they exist at all.

Today, I take a look behind the curtain of the most important committee that most people have never heard of, yet has influence that impacts each of us.

The CMIP scenarios are developed by a community of integrated assessment modellers numbering perhaps two hundred people worldwide, working in roughly fifteen institutions, and concentrated heavily in two: the International Institute for Applied Systems Analysis (IIASA) outside Vienna, and the Potsdam Institute for Climate Impact Research (PIK) in Germany.

Last month, this community released the illustrative versions of scenarios that will drive the next round of climate-model simulations that will be used in tens of thousands of research papers that project climate, climate impacts, economic consequences, and evaluate policy alternatives. These scenarios will underpin the next IPCC assessment due later this decade.

The illustrative scenarios are described a design paper in Geoscientific Model Development with 44 authors, and a follow-on Zenodo dataset lists 29 creators of that dataset.3 Sixteen of the 29 carry an IIASA or PIK affiliation.

Many of these researchers have been at the center of scenario development for two decades or more, spanning three iterations of scenario families.

I cross-referenced the 64 unique CMIP7 authors and dataset creators against the 13 papers that established earlier generations of scenarios: the Representative Concentration Pathways (RCPs) in 2011 and the Shared Socioeconomic Pathways (SSPs) in 2014 and 2017 yields:

• 21 of the 64 (33 percent) CMIP7 participants co-authored at least one foundational RCP or SSP paper.4

• Five participants co-authored all three foundational frameworks.

• Two of the three ScenarioMIP-CMIP7 co-chairs co-authored all three foundational frameworks.

The table below shows that those responsible for climate scenario generation have come from a very small group of individuals and institutions, and have maintained their role over multiple iterations of scenarios, especially the SSPs and CMIP7.

Two institutions in particular play an outsized role in scenario development: IIASA, the International Institute for Applied Systems Analysis in Austria, and PIK, the Potsdam Institute for Climate Impact Research, in Germany.

• IIASA, founded in 1972 as a Cold-War scientific-cooperation initiative between the United States and the Soviet Union, now operates as an international research institute funded by the scientific organisations of its 23 member countries.5 Its 2024 budget was €32 million, a little under 40 percent from member contributions. Since July 2023 it has been led by Hans Joachim Schellnhuber, who founded PIK in 1992 and ran it for a quarter-century before moving to Vienna to lead IIASA.

• PIK is a Leibniz Association member, funded jointly by the German federal government and the German states. In 2024 its institutional funding was €14.8 million; external project funding (largely European Commission) exceeded €25 million. It is led by economist Ottmar Edenhofer (former IPCC Working Group III co-chair) and earth-system scientist Johan Rockström. Since November 2024 PIK has been a registered entity in the Bundestag’s Lobby Register — which means that in addition to being a research institute, PIK is also a lobbying group that seeks to influence German government policy.6

The leaders of IIASA and PIK are collaborators in a shared political project. Schellnhuber directed PIK for 26 years (1992–2018) and was succeeded by economist Ottmar Edenhofer and Earth-system scientist Johan Rockström as co-directors in late September 2018. Both Schellnhuber and Rockström are co-authors of the 2009 paper that introduced “planetary-boundaries” — the framework that proposes a "safe operating space for humanity" defined by quantitative thresholds beyond which Earth-system change is held to be catastrophic. Rockström has explicitly connected the “planetary boundaries” concept to the neo-Malthusian ideas of Dennis and Donella Meadows of the Club of Rome (Limits to Growth).

The same individuals who lead the institutions that sit at the center of the production of CMIP7 marker scenarios are also among the leading public proponents of a very particular policy-relevant framing of what those scenarios should imply in policy.

The figure below shows the primary and secondary affiliations of the authors of the new paper describing the new CMIP scenarios (left) and the creators of the new scenario dataset (right). You can see the outsized role of IIASA and PIK.

Not only do IIASA and PIK play a central role in CMIP and IPCC, but they also have received substantial funding from climate advocacy philanthropy — ClimateWorks and Bloomberg Philanthropies. This funding supports the development and promotion of scenarios that are central to the regulation of central banks and businesses around the world — under the Network for Greening the Financial System (NGFS),7 a non-governmental consortium of central banks.8

• ClimateWorks Foundation is a San Francisco–based regranting hub founded in 2008 by the Hewlett, Packard, and McKnight foundations. Their goal is to keep warming to 1.5°C by directing philanthropic capital to the highest-leverage sectors and geographies for emissions reduction. It has awarded over $1 billion to more than 500 organizations since its founding, and was the sole philanthropic funder acknowledged in the NGFS Phase V Technical Documentation.

• Bloomberg Philanthropies is among the largest US private funders of climate advocacy. Its flagship initiative, Beyond Carbon — launched in 2019 with $500 million and doubled to $1 billion in 2023 — aims to retire every US coal plant, halve US natural gas capacity, block all new gas plants, and making “clean energy” 80 percent of US electricity generation by 2030, pursued through litigation, state-level lobbying, and grassroots advocacy. Michael Bloomberg also serves as the UN Secretary-General’s Special Envoy on Climate Ambition and Solutions. Bloomberg Philanthropies funded NGFS scenarios in Phases 1–3 but was not acknowledged in Phase V.

The Bloomberg and ClimateWorks funding of NGFS has downstream impacts on virtually every financial institution on the planet. An IIASA researcher — who has played a role in the development of the RCPs, SSPs, and CMIP7 scenarios — explains:

“Central banks obligate commercial banks to use our scenarios for climate financial stress testing and to assess the stability of the financial system against climate transition risk and climate physical risk. These scenarios will help banks make the economy more resistant to climate stress and navigate the financial sector toward the green transition.”

The lead modellers in creating the NGFS scenarios are also lead authors of the CMIP7 scenarios, and are funded by advocacy organizations with a stated interest in specific climate policies, with a rack record in promoting extreme, implausible scenarios. If that sounds odd or troubling — It is!

These cozy relationships among a very small group of institutions and individuals coupled with their outsized influence in global science, finance, amd policymaking raise a lot of questions.

Let’s look closer.

The figure below shows that across the new CMIP7 scenario paper and dataset, 36 unique institutions appear as primary affiliations.

• Fifteen are in Europe (42 percent of the total).

• Six are in the United States, the only North American country represented (another 17 percent).

• Five are elsewhere in Asia (Japan and Pakistan);

• Two are in India;

• One is in China; three in South America; two in Africa; two in Oceania.

Sixteen of 29 listed creators of the CMIP7 scenario dataset are at IIASA or PIK.9

Other researchers have raised questions about the concentration of climate scenario developers in European and North American institutions.

Tejal Kanitkar and colleagues showed in their 2024 Climate Policy analysis of 556 IPCC AR6 mitigation scenarios, the assumptions embedded in the IPCC scenarios systematically perpetuate and even expand North-South inequality.10 Across every scenario examined, per-capita GDP in regions containing 60 percent of the world’s population — Sub-Saharan Africa, South Asia, West Asia, and the rest of Asia — remains below the global average even in 2050.

They argue in that paper:

“A large majority of scenarios that are finally assessed in the IPCC 6th Assessment Report are submitted by modelling teams based in the global North… the pervasiveness of the absence of equity raises serious questions about the lack of diversity in the model building community, including the absence of perspectives from the global South.”

In the CMIP7 process China — responsible for over a quarter of global emissions — has just one institution represented. India, the world’s most populous country, has two. All of Africa has two. The communities whose long-run development paths are most at stake have the least input into the scenarios that describe those paths.

I am puzzled why the governments of these countries and regions have not already acted in response to their lack of input and influence on these incredibly important scientific advisory processes with broad global influence.

Consider typical expectations for any expert advisory body — much less one whose outputs move trillions of dollars and frame global energy and climate policies. Among those expectations:

• published Terms of Reference;

• a transparent procedure for selecting members;

• membership representative of nations and interests impacted by the expert advice;

• clearly stated criteria for how scenarios are created, included or excluded;

• an external peer community of comparable expertise to assess the choices made;

• opportunity for public comment on proposed products;

• conflict-of-interest disclosure and management.

I can locate no public documents from CMIP describing procedures for meeting these very basic expectations of good governance of institutions that provide influential scientific advice to policy makers. If they exist, they are well hidden.

The CMIP Panel itself appears to have acknowledged the good governance gap in its March 2025 report to the World Climate Research Programme’s Joint Scientific Committee:

“With the broadening of involvement across the CMIP project there is a need for more consistent and granular governance and guidance… CMIP is keen to discuss securing a mandate and discussing governance options for the informal consortium producing these datasets.”11

To put the governance gap in context, consider how comparable expert advisory bodies are governed.

• In the United States, the Federal Advisory Committee Act of 1972 (FACA) governs panels that give advice to federal agencies — bodies like the EPA's Science Advisory Board or the FDA's drug-approval panels. Under FACA, every covered committee must file a public charter spelling out its mission and membership rules; it must publish notice in the Federal Register at least fifteen days before each meeting; meetings must generally be open to the public, with members of the public allowed to file written statements and, at the agency's discretion, to address the committee directly; transcripts and working papers must be made available; and the committee's membership must be "fairly balanced in terms of the points of view represented."

• In the European Union, the European Food Safety Authority (EFSA), which assesses the safety of food additives, pesticides, and GMOs and whose opinions feed directly into Commission regulatory decisions, requires every expert on its panels and working groups to file an annual Declaration of Interests covering financial holdings, employment, consultancies, and research funding for themselves and their immediate families over the preceding five years; EFSA's Legal Affairs office screens each declaration for conflicts before each meeting, and experts with active commercial ties to the affected industries are excluded outright.

• The World Health Organization requires the same of every member of its expert panels, advisory committees, and guideline development groups; declarations are reviewed before appointment and again before each meeting, summary information is published with the resulting reports, and the WHO can require recusal or termination if a relevant interest emerges. Each of these regimes covers bodies whose individual decisions are typically narrower in consequence than the marker emission scenarios that anchor the next IPCC assessment, the climate stress tests run by every commercial bank, and national regulatory analyses.

The contrast between the governance of these important science advisory organizations with how CMIP scenarios are produced — without a public charter, without published Terms of Reference for the body that selects them, without standardized declarations of interest, with no transparency on how participating experts are selected, and with no opportunity for public input to proposed products — indicates that institutions of climate research created to inform policymakers needs to significantly improvements in their governance.

Comments, questions, suggestions, critique, debate, discussion — All welcome! Anyone from inside the processes or institutions that this post discusses is welcome to participate and share views, agree or disagree. Don’t want to be publicly acknowledged? No worries — send me a confidential email and you will remain anonymous.

Leave a comment

Before you go: The death of RCP8.5 and the institutions discussed today are being ignored by the “climate beat.” No matter — Help get the word out by clicking that “❤️ Like” button. More likes means more readers of THB. Thanks!

Subscribe now

Subscribe now

A look back at THB in April 2026.

It Really Happened

Following my post, there was lots of discussion and no doubt there will be much more. Here are a few:

• Michael Liebreich — RCP 8.5 is officially bollox

• Andy Revkin — The Old Worst-Case CO₂ Emissions Track Has Long Been History; Now RCP8.5 is Officially R.I.P.

• Zeke Hausfather — The arc of the scenario universe is long, but it bends inevitably toward more realistic emissions

Leave a comment

Climate Economics

I have been happy to see considerable discussion of the Curtin/Burgess economics paper, almost all of it very positive.

In the Wall Street Journal last week, I had an op-ed on climate economics, RCP8.5, and the ICAT “dataset” scandal — comments welcome!

• You Can’t Trust ‘Climate Economics’: Governments, banks and other institutions have based policies on models unconnected to reality.

Leave a comment

Data Updates and a Look Back at AIT

I updated our normalized U.S. tornado loss dataset through 2025 — extending and confirming our research from 15 years ago. I also took a deep dive into the economic consequences of the Iran war, which I’ll update again before long. And I took a look back at Al Gore’s movie from 2006, An Inconvenient Truth.

I did a couple of podcasts on Gore’s legacy. Here are links:

• RealClearPolitics: Al Gore and the Politicization of Science

• Science from the Fringe: ‘An Inconvenient Truth’ - 20 Years Later

Also, The Free Press published a version of my THB post: How Al Gore Politicized Climate Science.

Subscribe now

THB Follow-Ups and Series

I added significantly to several ongoing explorations here at THB: the widening FJC scandal and Part 3 of my series on the economic weight of energy.

You can read my initial post on the FJC scandal here: Who Actually Wrote the Climate Manual for Federal Judges.

This post is Part 3 in the THB series on the shrinking economic weight of energy which appeared as my monthly column in The Dispatch. The first two installments in the series are:

• Part 1: A Fossil Fuel Energy Transition Started 50 Years Ago

• Part 2: How can fossil-free energy sources outcompete fossil fuels?

Leave a comment

The First THB Deep Dive Report and a new Five Figures

I produced the first THB deep dive report, on universities. I’ve been hard at work on the second deep dive, which will be on the El Niño-Southern Oscillation (ENSO), due to expectations that the world is heading into a (perhaps large) El Niño event. I’ve decided that the deep dives will be moved out from behind the THB paywall one month after publication, so paid subscribers get the first look.

I typically produce much more work than I can share here at THB, and Five Figures gives me a chance to share some of that work. The topics span a large range of things I’m interested in or encounter in my work.

Leave a comment

A Few Announcements

• I have updated the THB Series page, which groups THB posts together that explore a theme over time. Let me know your suggestions for how I might improve this collection.

• I have also updated the For Paid Subscribers page, which offers one-stop access to PDFs of three of my books (so far!) — Hurricanes (with Pielke Sr.), Disasters and Climate Change, and The Climate Fix. Also available are paywalled posts — THB Insider, Five Figures, and those that share various otherwise paywalled PDFs of my published peer-reviewed work. More updates in the works.

• I’m working on the 3rd edition of Disasters and Climate Change — I’m not sure if I’ll look for a publisher (ASU/CSPO no longer does their short-book series in which the first two editions were published) or I’ll just self-publish here at THB.

• Note: The annual subscription for new subscribers to THB is increasing to $100 on 1 June 2026. If you have been on the fence about supporting THB, now is the best time to lock in your support at a lower price that will never increase!

Comments, critique, discussion — All welcome!

Leave a comment

Before you go — If you find value in this monthly compendium, please let me and everyone else know by clicking that “❤️ Like.” Thanks!

Share The Honest Broker

Subscribe now

The international committee responsible for the official scenarios that feed into climate modeling that are the basis for most projective climate research and the assessments of the Intergovernmental Panel on Climate Change (IPCC) has just published the next generation of climate scenarios.

Big news: The new framework has eliminated the most extreme scenarios that have dominated climate research over much of the past several decades — specifically, RCP8.5, SSP5-8.5, and SSP3-7.0. This is an absolutely huge development in climate science which will have lasting impacts across research and policy.

The future is not what it used to be.

Today’s post commends the researchers who have brought climate scenarios more in line with current understandings, but also raises some significant continuing issues with the scenarios.

Let’s get started . . .

The new scenarios come from the Coupled Model Intercomparison Project (CMIP) — a project of the World Climate Research Programme (WCRP), co-sponsored by the World Meteorological Organization, the International Science Council, and UNESCO’s Intergovernmental Oceanographic Commission.

Under CMIP, now in its seventh iteration, sits another little-known committee with responsibility for developing the scenarios necessary for earth system models to project future climate.1 That committee — called ScenarioMIP — just published the new scenario framework that will underpin the IPCC’s Seventh Assessment Report (AR7) and much of the research that it will draw upon.

In a paper released earlier this month, Van Vuuren et al. (VVetal26) introduce a new set of seven scenarios. The authors write of the obsolete high end emissions scenarios (emphasis added):

“For the 21st century, this range will be smaller than assessed before: on the high-end of the range, the CMIP6 high emission levels (quantified by SSP5-8.5) have become implausible, based on trends in the costs of renewables, the emergence of climate policy and recent emission trends.”

Read that again — The high end scenarios are Implausible.2

I disagree that the implausibility of the high-end scenarios resulted from the falling costs of renewables or the emergence of climate policy, but that is a debate for another day.

What matters today is that the group with official responsibility for developing climate scenarios for the IPCC and broader research community has now admitted that the scenarios that have dominated climate research, assessment, and policy during the past two cycles of the IPCC assessment process are implausible: They describe impossible futures.

Tens of thousands of research papers have been — and continue to be — published using these scenarios, a similar number of media headlines have amplified their findings, and governments and international organization have built these implausible scenarios into policy and regulation.

We now know that all of this is built on a foundation of sand.

What changed

The new CMIP7 ScenarioMIP framework offers seven scenarios spanning a range from “VERY LOW” through “HIGH.” The current naming convention drops the radiative-forcing target labels of the SSP era — there is no “8.5” scenario, and no “7.0” scenario, but as I’ll show below, each scenario has a radiative forcing level in 2100.

I ran the available new scenarios (HIGH, MEDIUM, LOW, and VERY LOW) through the FaIR calibrated and constrained ensemble that Sanderson and Smith (2025) used to characterize the CMIP7 set (FaIR v. 2.2.0 as described in their README file). I then ran each of the five tier-1 SSPs through the same emulator with identical parameters to ensure that the results are apples-to-apples. The full methodology, data, and code is in the appendix to this post.

The headline results follow.

CO₂ emissions: fossil fuels and industry, 2000–2100

The chart above shows fossil-fuel and industry CO₂ emissions for four CMIP7 scenarios alongside the five tier-1 SSPs and the two main reference scenarios from the 2025 IEA World Energy Outlook.

Note the massive gap between the new HIGH and SSP5-8.5. The new HIGH reaches 71 Gt CO₂/yr in 2100 — far below SSP5-8.5 at 128 Gt in 2100. Nothing in the CMIP7 set comes close to SSP5-8.5. The new HIGH also sits below SSP3-7.0 by about 9% in terms of cumulative emissions to 2100. Note also the gap between MEDIUM (solid yellow) and SSP2-4.5 (dashed yellow), which I’ll return to below.

Both of the most recent IEA near term scenarios — which run to 2050 — fall below MEDIUM and SSP2-4.5.

The table below compares the CMIP7 scenarios to their closest AR6 analogues, showing that the overall range has constricted. The higher scenarios have come down and the lower scenarios have come up — except VERY LOW, which moved down.

2100 effective radiative forcing and end-of-century temperature

The table below lists AR6 and CMIP7 scenarios from highest to lowest 2100 radiative forcing. The middle column shows the average global temperature change from an 1850-1900 baseline, under the climate emulator used by CMIP7. The right column shows the average temperature change for the SSPs as projected by the IPCC AR6.

Interestingly, the projected 2080-2100 temperatures of the SSPs decreased from their AR6 values based solely on recent updates to the FaIR climate emulator.3 These changes resulted primarily from the updating of emissions trajectories from 2014 (used in AR6) to 2023 (used by CMIP7). The more moderate emissions trajectories resulted in lower projected end-of-century temperature increases.

The new CMIP7 HIGH is 0.9°C cooler than SSP5-8.5 in apples-to-apples terms (and 1.4°C cooler versus IPCC AR6), and 0.2°C cooler than SSP3-7.0 (-0.6°C against IPCC AR6).

The implausibility of upper-end legacy scenarios is now official.

CMIP7 avoided repeating the past with SSP3-7.0

Last April I argued here at THB that the climate science community was on the brink of repeating the RCP8.5 mistake with SSP3-7.0 — which assumed a 2100 global population approaching 13 billion, well above any contemporary demographic projection and a five-fold expansion of global coal use. Neither assumption survives current understandings of demographics or energy systems.

I don’t know if anyone in CMIP or ScenarioMIP reads THB4 — if not they should! — but regardless, they wisely chose not to adopt SSP3-7.0 as the new HIGH scenario.

The new HIGH scenario sits at 6.7 W/m² in 2100 — below the SSP3 baseline 7.0 W/m² — with 9 percent less cumulative fossil CO₂ through 2100. As I’ll discuss below, this is progress; it is partial but real.

But the new HIGH still sits well above the plausibility range that we identified in Pielke, Burgess, and Ritchie (2022). We found that of the >1,000 scenarios in the AR5 database, the plausible subset centered on a median of ~3.4 W/m² in 2100, with an upper end near 6 W/m². The new HIGH scenario is well above that upper end.

The authors of Van Vuuren et al. partially acknowledge that the new HIGH scenario is exploratory — a thought experiment, not a projection:

“Clearly, this scenario is not a “business-as-usual” scenario nor the no-policy reference scenario for the other scenarios. The scenario is intended to explore the upper end of GHG emissions resulting from deep political, technological, and structural deviation from current trends.”

Note that first sentence — It means that any future research that compares the HIGH scenario to lower scenarios in order to characterize the effects of climate policy will be fundamentally flawed. The HIGH scenario is not a projective scenario, but a “what if?” exercise.

Unfortunately, Van Vuuren et al. then engage in some unsupported speculation about the plausibility of the HIGH scenario in the real world:

“The are various reasons why such a scenario could emerge. For instance, a rollback of climate policies could result from a lack of public support for the energy transition. This could be related to, for instance, local opposition to building new wind farms or concerns about impacts on fossil industries related to jobs and national energy security. Also, the rapid cost decrease in renewable energy of the past decade could be discontinued, possibly as a result of regional scarcity and limited tradability in materials for solar and wind technologies and EV batteries . . .”

As discussed below, the new population assumptions of the updated SSP3 are ridiculous, and by themselves render the HIGH scenario implausible. The lack of any systematic effort to evaluate plausibility of scenarios remains a fundamental weakness of the scenario development process.

The new scenarios are SSPs in new clothing

The new CMIP7 framework does not start from a fresh socioeconomic foundation. Van Vuuren et al. explain:

“In practice, the IAM [Integrated Assessment Modeling] teams have based their current scenarios on various SSPs, as it was generally deemed pragmatic as these come with already available, suitably rich quantifications and were implemented by the participating modelling teams within the given timeline.”

The CMIP7 scenarios rely on the same narrative architecture as the IPCC AR6 SSPs. The table below shows how the SSP storylines map onto the new scenarios — Confusion is sure to result, as the old SSPs are not the new SSPs.

The new HIGH inherits the SSP3 storyline directly — the same SSP3 whose enormous 2100 population rendered it implausible when used in the IPCC AR6. Remarkably, the IIASA 2024 update to the SSPs did not bring the population projections down. Instead, it increased them, as you can see in the table below.

The 2024 demographic update (KC et al. 2024, IIASA Working Paper WP-24-003, also referred to as WIC2023) revises SSP 2100 populations upward: SSP3 jumps from 12.6 to 14.5 billion. SSP4 increases from 9.3 to 13.3 billion — an eye-popping 43% increase.

The SSP population update assumes that child mortality declines faster than WIC2013 anticipated, fertility declines slower in sub-Saharan Africa, and Africa’s 2020 base-year population was already 76 million higher than WIC2013 had projected. Africa’s 2100 population alone is now 3.55 billion, up from 2.62 billion in the AR6-era projection — a 35% upward revision.

This puts the CMIP7 HIGH scenario in a strange position: The cumulative fossil CO₂ from energy and industry in CMIP7 HIGH (4,629 Gt 2020-2100) is lower than SSP3-7.0 (5,074 Gt), but it has a 2100 population that is 15 percent larger. That means that the implied per-capita emissions trajectory in the new HIGH has a steeper decline.

How much of the new HIGH scenario’s warming is due to its incredible population projection?

I performed a simple sensitivity analysis:

• Take HIGH as published: 4,629 Gt cumulative fossil CO₂ over 2020-2100, producing about 3.0°C of 2081-2100 warming.

• That means per-capita emissions intensity average about 5.2 tonnes of CO₂ per person per year using 11-billion as the mean population across the century.

• Hold that per-capita intensity constant and replace the SSP3 population trajectory with the SSP1/SSP5 population trajectory — peaking at 8.5 billion in 2050, declining to 7.4 billion by 2100, averaging about 8 billion.

Under these assumptions, under HIGH, cumulative fossil CO₂ falls to roughly 3,330 Gt, a 1,300 Gt reduction. The IPCC AR6 central TCRE estimate (0.45°C per 1,000 Gt CO₂) gives about 0.6°C less warming. The HIGH scenario with SSP1/SSP5 population thus would deliver about 2.4°C of 2081-2100 warming — slightly cooler than the new MEDIUM at 2.5°C.

Under this simple method, about 0.6°C of the HIGH scenario’s projected warming traces to the population assumption alone. The HIGH scenario may be much less about carbon and much more about assumed human fecundity.

This is a sensitivity analysis, not a coherent scenario — combining SSP3’s per-capita intensity with the SSP1/SSP5 demographic profile is internally inconsistent. But it suggests that the demographic contribution to the HIGH scenario’s warming is significant.

The plausibility vacuum remains

The deeper problem with the SSP/RCP architecture, as Justin Ritchie has documented at length, is that physical climate modeling became decoupled from the underlying IAM socio-economic scenarios.

Under the RCPs, scenario creators identified concentration pathways and the underlying socio-economic assumptions were expected to be filled in later. Whether the underlying assumptions actually described a coherent picture of the world was never systematically assessed.

Ritchie called this a plausibility vacuum — a situation where any combination of climate model inputs could be used without any assessment of the real-world plausibility of the assumptions.

To be fair, the new CMIP7 framework does address some earlier shortfalls. The new design specifies emission-driven runs as the default, which allows for carbon-cycle feedbacks. The harmonization of emissions to observed 2023 data is an improvement — CMIP6 harmonized to 2014, and that harmonization had become well out-of-date by the time AR6 came out.

However, the plausibility vacuum problem remains. Van Vuuren et al. do not evaluate scenario plausibility against observed energy trends, against IEA projections, or against the body of literature critiquing the SSP set.

The 2023 ScenarioMIP workshop report — which kicked off the process of developing these new scenarios — recognized the plausibility problem and committed to addressing it. Van Vuuren et al. does not deliver on that commitment.

ScenarioMIP produces a better scenario set than those of AR6, but the improvement comes from incorporating more recent emissions data and accepting the undeniable collapse of SSP5-8.5’s credibility — not from any methodological reform of how scenario plausibility is assessed.

The MEDIUM scenario is not “current policy”

The walk-back at the high end is the most important change in the new framework. The story in the middle is more complicated.

Van Vuuren et al. describe the MEDIUM scenario as one that “shows the consequences of the current policy situation (as of 2025) and trends continuing over the century.” They specify that MEDIUM includes only policies “actually officially being implemented” — no NDC pledges, no net-zero announcements, unless backed by explicit policy. The framing implies that MEDIUM tracks where the world is actually headed under current policies.

That framing is not consistent with other approaches to defining a current policy trajectory. The CMIP7 MEDIUM scenario produces fossil-fuel CO₂ emissions that rise from about 38 Gt CO₂/yr today to 41 Gt by 2050 — that part compares well to the scenarios (CPS and STEPS) of IEA’s World Energy Outlook 2025, released last November.

However, it is post-2050 where the new MEDIUM scenario diverges from other current-policy projections, with its emissions slowly rising through the second half of the century.

In contrast, STEPS — the IEA’s standard reference for where current and announced policies actually take us — falls below 30 Gt by 2050 and produces about 2.5°C of 2100 warming.

The CMIP7 MEDIUM is more accurately characterized as a “policy stagnation” scenario rather than a “current policy” scenario. Cumulative fossil CO₂ emissions through the end of the century come in 18% higher than SSP2-4.5 — the AR6-era middle-of-the-road scenario — even though the MEDIUM 2081–2100 mean temperature is slightly cooler (explained below).

The figure below shows what happens once SSP2-4.5 and MEDIUM are anchored to the same observed 2020 baseline. The original SSP2-4.5 (black) sits roughly 0.6°C above CMIP7 MEDIUM (red) at 2020 — not because the 21st century plays out differently, but because the SSP framework's historical emissions are harmonized to 2014 while CMIP7's are harmonized to 2023. That head start propagates through the entire 21st century trajectory.

Running SSP2-4.5 post-2020 emissions through the CMIP7 FaIR emulator with CMIP7’s updated historical baseline (dashed blue), the picture inverts: SSP2-4.5 produces a 2081-2100 mean of 2.44°C against CMIP7 MEDIUM 2.56°C. On an apples-to-apples comparison, the new MEDIUM results in about 0.12-0.20°C more warming than SSP2-4.5 — due to its 520 Gt larger cumulative CO₂ budget over the century.

This connects to the larger argument in Pielke, Burgess, and Ritchie (2022): observed CO₂ emissions from 2005 through 2020 tracked closer to the SSP 3.4 W/m² range than to SSP2-4.5. Today, the gap has widened, not closed. IEA STEPS now has emissions falling to under 30 Gt by 2050 — a trajectory consistent with the SSP 3.4 forcing range. The center of the new CMIP7 scenario set sits well above that range.

The new framework compressed the high end. Good.

The middle did not move far enough. The new MEDIUM might be considered a worst-case scenario rather than a current policy scenario. Interestingly, that would mean that there is no real current-policy scenario in the new CMIP7 scenarios, which would be something like a SSP2-3.4 scenario with updated demographics.

Perhaps, if projected demographic changes continue to trend lower, CMIP7 LOW will come to represent something akin to a current policy trajectory.

Why this matters: these scenarios live in policy

The now-implausible upper-end scenarios — RCP8.5, SSP5-8.5, and SSP3-7.0 — are not just academic constructs used in esoteric research. They are embedded in the policies and regulations of most of the world’s largest economies, found across the world’s most important multilateral institutions, and used in the climate stress tests that govern hundreds of billions of dollars in bank capital.

The table below provides just a few examples.

National climate impact assessments in the United States, United Kingdom, Germany, Canada, Australia, Japan, and the Netherlands all use RCP8.5 or SSP5-8.5 as a reference scenario. The Network for Greening the Financial System framework, used by more than 140 central banks, has utilized a “Hot House World” scenario calibrated to RCP8.5 physical risk into the bank stress tests run by the European Central Bank, the Bank of England, the Reserve Bank of New Zealand, the Banque de France, and the US Federal Reserve. The World Bank’s Climate Change Knowledge Portal, which provides the climate diagnostics that feed into the Country Climate and Development Reports for more than 100 client countries, defaults to SSP5-8.5 and SSP3-7.0.

The abandonment of the high-end legacy scenarios by CMIP7 will need to propagate through this entire infrastructure. The policy machinery built on RCP8.5 and the other implausible scenarios is systemic.

What this means

The new CMIP7 ScenarioMIP framework represents a real course correction, but with more work to do. SSP5-8.5 is gone. SSP3-7.0 has a successor that is less extreme but arguably remains implausible. The middle of the set is more pessimistic than trajectories of current and announced policies. The plausibility vacuum at the heart of the architecture has yet to be addressed.

All this means that users of climate models and model output based on legacy scenarios will now face decisions about if and how they’d like to realign with the latest scientific understandings versus continuing to rely on outdated research.

Furthermore, there are no doubt many — hundreds if not thousands — of studies in the publication pipeline that depend upon the upper end scenarios. Editors and reviewers should ensure that they are properly characterized as exploratory and are not intended to be interpreted as projective.

We’ve known since 2017 that upper end climate scenarios are fatally flawed. Nine years later, that understanding has now become officially recognized. That is good news.

We can debate whether nine years is short or long for the overturning of scientific understandings with massive economic and policy implications. But today, that overturning is undeniable.

Science is self-correcting. What matters now is what happens next.

Comments, questions, suggestions, critique, debate, discussion — All welcome!

Leave a comment

Before you go: Please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

Subscribe now

Methods Appendix

The quantitative analyses in this post and production of figures and tables were performed under my direction by Anthropic’s Claude Optus 4.7. A discussion of methods can be found in the PDF below. After the jump, paid subscribers can access an Excel spreadsheet with all of the data in the figures and tables in this post and a Python script that will reproduce all computations in this post. This is work in progress. Comments welcome!

Thb Cmip7 Methods Appendix (2)

3.27MB ∙ PDF file

Download

Download

Read more

Subscribe now

At a moment of heightened polarization on nearly every issue facing the nation, it’s no wonder that energy consumption and climate change have become a political football for U.S. lawmakers left and right. But are Americans really as divided on these key public policy problems as they appear at first glance? A simple principle may offer us an answer.

The iron law of climate policy holds that when politicians force a choice between economic growth and emissions reductions, economic growth wins. Not because politicians are cynical or voters don’t care about the environment—most do—but because energy is foundational to everything else: to jobs, to heating and cooling homes, to food, to medicine, to the entire material standard of living. The iron law is every bit as real as the physics of energy and climate.

That reality shows up clearly in public opinion data. Just before the 2024 election, my American Enterprise Institute colleague Ruy Teixeira and I conducted a survey of what Americans think about energy and climate, a YouGov poll of more than 3,000 registered voters.

One result stood out above all others: An “all-of-the-above” approach to energy policy—one that includes oil, gas, solar, wind, and nuclear—commands majority support across every demographic and political group that we measured. It is the only question in the entire survey where we saw that degree of agreement, a rarity given today’s polarization on just about everything.

The survey also reveals that voters are ruthlessly pragmatic about energy costs. Asked whether they would support a $1 monthly fee on their electricity bill to fight climate change, only 47 percent of respondents said yes and 43 percent said no. Raise the fee to $20 and support collapses to 26 percent. At $100, the ratio is 7-to-1 against. Working-class voters are even more resistant to such fees at every level. Voters also ranked “dealing with global climate change” 15^th out of 18 policy priorities—well behind keeping energy costs low, fighting inflation, and strengthening the economy.

The survey’s deeper message: Voters want lower prices and secure supply, and they want them across all energy sources. They generally support action on climate change, but only after these priorities are met. And while the U.S. energy supply and demand landscape is varied, there are issues on which almost all Americans agree. Politicians of both parties should take note.

To navigate the politics of energy, it is helpful to understand the notion of energy intensity: How much energy does an economy consume per dollar of output? If the economy grows faster than energy use, intensity falls. If energy use grows faster, intensity rises. The Kaya identity—the standard framework in climate policy for decomposing carbon emissions—treats energy intensity as one of four levers of decarbonization, alongside population, per-capita GDP, and the carbon intensity of energy supply.

There are two ways to measure intensity. Physical intensity tracks how much raw energy—barrels of gasoline, kilowatt-hours of electricity, tons of coal—flows through the economy per dollar of real output. Expenditure intensity tracks how much money Americans spend on energy as a share of GDP. Physical intensity tells you about efficiency and economic structure. Expenditure intensity tells you about the economic burden—which, as the iron law explains, is what drives the politics.

In two recent analyses for The Honest Broker on Substack, I looked at energy intensity across eight fuel and generation sources—gasoline, natural gas, coal, retail electricity, nuclear, hydro, wind, and solar—from 1970 to 2024. During this period, real U.S. GDP grew 341 percent while total primary energy consumption grew 39 percent, meaning that the economy grew roughly nine times faster than energy use. These technical details help us to better understand the bigger picture of the politics of energy in America.

A key reason for the gap between economic growth and increasing energy consumption was the oil shocks of 1973 and 1979, which led to a period of prolonged high energy prices that touched the entire economy. That signal drove policy-motivated and other efficiency investments in vehicles, buildings, and industrial processes that compounded for decades. At the same time, the composition of American output shifted toward less energy-intensive activities. Services, software, finance, and health care now occupy a larger share of GDP than they did 50 years ago, and they run on far less energy per dollar than the manufacturing and heavy industry they partially displaced.

Policies that work with that prevailing dynamic, accelerating efficiency and innovation while keeping energy affordable, enjoy broad support. Policies that make energy more expensive, either as an unintended consequence or intentionally to coerce changes in supply or demand, tend to generate a political backlash—such as the backlash we see today in widespread public concern about affordability and inflation.

To better understand the political geography of energy, I compiled state-level data on electricity consumption, motor gasoline consumption, gasoline prices, population, and real GDP in 2023. These data tell two interesting stories.

Electricity intensity differs in red vs. blue states

As the graph below shows, there is a strong correlation between a state’s electricity consumption per $1,000 GDP and its Trump vote share in the 2024 election. Republican-leaning states include the major energy producers—such as Texas, Wyoming, West Virginia, and North Dakota—and states with large industrial and agricultural economies. For instance, Wyoming’s physical electricity intensity is roughly six times that of California. Energy-intensive activities cluster in red states, while services and knowledge industries cluster more in blue states.

But gasoline expenditure is a bipartisan story

When we shift from physical electricity intensity to what consumers actually spend on gasoline, the correlation nearly disappears. Interestingly, consumers in high-consumption red states and low-consumption blue states end up spending similar amounts on gasoline, due to differences in gasoline prices. Red states consume more gallons per capita but pay less at the pump—due to lower state taxes, shorter supply chains, and the absence of California-style fuel blend requirements. Blue states consume less but pay more per gallon.

In practice, the fact that everyone feels the squeeze of higher gasoline prices means politicians from both parties will face backlash, and that backlash falls disproportionately on whoever is in power, regardless of the underlying causes. The iron law does not discriminate.

American spending on gasoline, natural gas, coal, and retail electricity together totaled roughly 3.8 percent of the real GDP in 2024. At its peak in the early 1980s, direct expenditures on those fuel sources represented more than 12 percent of national GDP, which helps to explain why the energy shocks of the 1970s continue to reverberate through the U.S. energy economy 50 years later.

At the same time, the United States devotes a smaller fraction of national income to direct energy expenditures today than at any point in the past half-century. But it is the relative change in that burden that matters most in today’s politics—not the historical context. A move from 3.8 percent to 5 percent of GDP would be politically significant even if 5 percent remains historically low. It is not the absolute level of the energy burden that drives political behavior, but the direction of change and who bears its cost. Lower prices and security of supply are shared priorities across the partisan divide.

Support for an all-of-the-above approach—which prioritizes both affordability and diversity of supply—is the only unanimous majority position across every demographic and political group in our survey. This preference reflects voters’ shared direct experience of energy as a foundational economic good. Fossil fuels, wind turbines, and solar panels are often held up as political symbols by one side or the other, but when prices go up, posturing and symbolism are quickly shed in favor of realism and pragmatism.

Whether they vote Republican or Democrat, Americans want the lights to stay on and transportation to be affordable, regardless of what source or technology is providing that energy. The partisan divide in energy—as sharp as it may look among politicians and pundits—is far narrower at the level of normal Americans, wherever they live and whomever they vote for.

Evidence shows that the U.S. economy has been decarbonizing, steadily, for half a century. The mechanisms that drove that decarbonization are efficiency compounding through market signals, structural economic change, and targeted technological innovation driven by private and public sectors working together. If decarbonization of the economy is to continue in an era of growing energy demand—which it should, because it helps to drive the economic growth that benefits citizens and consumers—then these mechanisms will need to be sustained through conscious, collaborative effort.

The iron law is not an obstacle to effective energy or climate policies. Instead, it offers a boundary condition for policy design—everything starts with affordable, reliable, abundant energy—to achieve goals that a strong majority of Americans, across party lines, already support. An all-of-the-above strategy that keeps energy cheap and secure, while steadily improving the efficiency and carbon intensity of the supply, is not an obstacle to the decarbonization agenda. It is the precondition for it.

Policy Watch

• A new survey of more than 3,500 U.S. political elites—including elected officials, staffers, regulators, lobbyists, and journalists—found that fewer than half of Republican respondents affirmed that global warming is primarily human-caused, with nearly one-third deeming a climate conspiracy claim—that “global warming is based on manipulated data to suppress dissent and undermine the economic dominance of the United States”—as more likely true than not. Democratic elites are near-unanimous in believing that warming is happening and due to human activities. Partisan identity explains more than 15 times as much variation in elite climate beliefs as general trust in scientists—more than ideology, conspiratorial disposition, education, or professional background. Efforts to use science to confront elites who dismiss anthropogenic climate change does not address the mechanism driving their dismissal—polarization, not data.

• Tatsuya Terazawa, the chairman of the Institute of Energy Economics, Japan, recently published the clearest analysis available of what the current Iran crisis looks like from Tokyo. Japan imports nearly all of its oil and liquefied natural gas, making it highly dependent on Middle East exporters. But it also took steps to mitigate the current shortages, as Terazawa explains: “Japan had amassed 254 days’ worth of reserves, including five months’ worth as SPR managed by the Government, by the time of the attack on Iran on February 28. The high level of reserves has been one of the major pillars of energy security policies in Japan.” But Japan has more work to do to shore up its energy security, he argues. “After the resolution of the conflict, we should not forget the vulnerabilities of our energy system that have been painfully exposed.”

Innovation Spotlight

• In a new report for the American Enterprise Institute, David Hart critiqued Vannevar Bush’s 1945 metaphor of an “endless frontier” for future scientific progress, an idea that has long shaped U.S. policy. When it comes to energy technology today, Hart concludes, the frontier is no longer open. Chinese advancements—in fields ranging from polysilicon and solar modules to battery chemistry and EV supply chains—have already settled much of it. American innovators contesting that terrain are not pioneers in open country; they are operating in a zone of conflict against entrenched incumbents with substantial first-mover advantages. The policy implication: The standard toolkit of basic research funding and startup support is insufficient for this environment. Hart advocates for a systematic deployment of fiscal, regulatory, informational, and diplomatic tools—but focused narrowly on a few high-priority domains where the U.S. has a genuine comparative advantage rather than attempting broad leadership across the energy technology landscape. Scope, he argues, is a strategic asset when institutional capacity is limited and political coalitions are fragile.

Further Reading

• Reuters energy columnist John Kemp just released his recommended reading for 2026. The bibliography covers production, consumption, transportation, technology, trading, economic history, and geography—and his criterion for inclusion is straightforward: He has read it and found it useful.

Comments welcome!

Leave a comment

Before you go: Please click that “❤️ Like” button. More likes means more readers of THB. Thanks!

Read more

Subscribe now

There has been so much to cover here at THB of late that I’ve neglected the Five Figures series. On the bright side, that means I’ve collected and produced some great stuff — and today it is actually 9 figures.

This month:

• Very interesting oil data: When will we again see $3 gasoline?1

• Colorado River record low flow

• Faculty ideology trends

• Unprecedented military firings

• Sports: Goat and unicorn

Let’s go…

Read more

Part 2 of 2.1 Part 1 explains the Curtin and Burgess (CB26) argument why the climate-economy relationship is empirically inscrutable. Here I walk through what Curtin and Burgess found when they replicated and stress-tested three of the most influential top-down climate damage studies — and what follows from their results.

Subscribe now

Last December, Nature retracted “The Economic Commitment of Climate Change,” by Kotz, Levermann, and Wenz (KLW24) — one of the most influential climate economics papers of the past decade. The paper claimed that climate change would cost the global economy $38 trillion a year by 2049 and projected an income reduction of 19 percent within 26 years regardless of future emissions.

KLW24 was the second most mentioned climate paper by the media in 2024, according to Carbon Brief. The paper was cited by central banks and governments to justify more aggressive climate policies.

I was among those who viewed the retraction as good news: science self-correcting, a bad paper removed, maybe things are getting back on track. It turns out there is more to the story — Much more.

The Kotz retraction was not a one-off case of flawed science belatedly retracted.

The new preprint by Finbar Curtin and Matt Burgess, of the University of Wyoming — “The empirically inscrutable climate-economy relationship” — makes undeniably clear that the KLW24 retraction was just a symptom of a much deeper problem in climate economics.

The methodological problems that result in KLW24 run through the entire top-down climate-economics literature. If it was appropriate to retract KLW24, then what now should happen to the hundreds of other papers with the same methodological shortfalls?

In Part 1, I explained the CB26 theoretical argument: the data structure economists use to estimate climate damages cannot, in principle, recover the thing it is trying to measure. Country-year panels pool observations across space and time where and when the relationship between temperature and economic output differs enormously — El Salvador is not Iraq, and India in 1970 is not India in 2020.

Fixed effects cannot solve that problem. The degrees of freedom run out. CB26 call that relationship empirically inscrutable. I go further and argue that seeking to connect two indicies — climate and GDP — can produce numbers, but cannot lead to meaningful results.

Today, in Part 2 I overview the CB26 replication of the three most influential papers in the top-down climate-economics literature. Each paper has similar weaknesses to those that resulted in the retraction of Kotz et al. 2024.

The table below summarizes what CB26 found with respect to those three papers: Burke, Hsiang, and Miguel 2015 (BHM15), Kahn et al. 2021 (KETAL21), and Bilal and Känzig 2025 (BK25).

Across these three papers are the same structural problems as KLW24:

• Results driven by a small number of outlier observations that have nothing to do with weather;

• Sensitivity to specification choices;

• And, the pooling of data across places and times where the relationship of interest clearly differs.

These are not minor technical quibbles. They are the core identification failures that CB26 identified theoretically and also find in practice..

The retraction of KLW24 was due to data errors and an arbitrary lag structure. The data errors were the proximate cause — a decimal-vs.-percentage-point mistake that reviewers should have caught.

The lag structure was arguably worse: Kotz et al. estimated that a year-to-year change in temperature carried effects equally large seven years later as in the present year, then added those lagged effects across years to produce damage totals that dwarfed everything in the prior literature. Remove a single anomalous observation — the Uzbekistan outlier — and the damage estimates collapsed.

Now look at the table above.

• BHM15 nonlinear damage function attenuates when a handful of “growth miracles and growth disasters” leave the sample — events like Iraq’s post-2003 rebound, the Soviet collapse, Oman’s oil boom. These are real economic events with nothing to do with temperature, but they happened to fall in years of unusual weather in countries with volatile climates, and the regression assigned the weather as cause.

• KETAL21 long-run growth effects vanish with different model specifications.

• BK25 $1,200-per-tonne social cost of carbon disappears when local temperatures replace a global average index.

The same methodological issues appear across all four papers: results driven by influential outlier observations, damage estimates sensitive to choices the authors present as defaults, and a fundamental inability to distinguish a real climate-economy signal from statistical artifacts produced by pooling incompatible data.

The Kotz retraction removed one flawed paper, leaving behind a flawed literature.

CB26 focused on three papers, but their critique applies to the entire literature: The top-down methodology those papers rely on — panel regressions of GDP, agricultural output, labor productivity, or mortality on temperature and precipitation, with country or region fixed effects, across multi-decade samples — runs through a large fraction of the empirical climate-economics literature published over the past fifteen years.

The literature CB26 critique includes Dell, Jones and Olken (2012), Moore and Diaz (2015), Kalkuhl and Wenz (2020), Nath, Ramey and Klenow (2024) — each

“relating an aggregate climate index to an aggregate economic index and interpreting the resulting coefficient as a causal damage function.”

The three papers CB26 replicate are the most prominent examples of an approach shared across hundreds of published studies.

The top-down literature did not stay in academic journals. Top-down climate economics papers have been profoundly influential in regulatory frameworks, financial supervisory standards, and legal arguments that govern trillions of dollars in decisions. The table below documents some examples where those studies showed up in finance and policy.2

Long-time THB readers will know that I have been writing for years about implausible climate scenarios — Specifically, how the high-end RCP8.5/SSP5-8.5 emissions scenario was the basis for tens of thousands of research papers, producing the most alarming results, and achieving enormous institutional standing. We now know that those extreme scearnios are implausible — and we have known that for a while, thanks to my colleague Justin Ritchie’s work from a decade ago.

The top-down climate-economics literature has followed a similar path. An approach disconnected from the real world in specific, identifiable ways achieved enormous real-world influence.

In a 2021 paper, Ritchie and I argued that RCP8.5 came to have such a tight grip on climate research due to a confluence of factors — including incentives faced by researchers, a world that welcomes apocalyptic scenarios, and the rewards of publishing research deemed supportive of climate action. I would guess that some combination of those factors are at play in climate economics as well.

For climate economics, I’d add another factor to the mix — climate determinism.

The core idea is old and repeatedly discredited: that climate conditions explain a large share of variation in human outcomes including prosperity, social organization, and economic development.

In the nineteenth century, scholars attributed tropical poverty to the heat. In the early twentieth century, Ellsworth Huntington built a career arguing that temperate climates produced superior civilizations. These ideas fell from favor because the evidence did not support them and the ideological freight they carried was obvious.

Climate determinism has returned — Instead of historians speculating about civilizational effects of climate, economists ran panel regressions with fixed effects and reported precise numerical estimates of how a one-degree rise in a global index of temperature reduces GDP.

The math was new. The underlying claim was old: climate drives economic outcomes in ways that dominate human adaptation and institutional capacity.

The economists working in this literature did not set out to resurrect climate determinism — I have no doubt that they all understand perfectly well that institutions, technology, and governance matter enormously.

However, the top-down methodology they employed requires a simplistic assumption that an aggregated index of weather over a year at the global level can explain outcomes across the global economy. That is climate determinism.

Curtin and Burgess explain the significance of their paper, which in my view is seminal:

Our analysis should provoke a fundamental reevaluation of how climate-econometric studies are used and referenced in research and policymaking. Estimates of economic damages from climate change—and other related quantities like the SCC—affect trillions of dollars in public and private decisions. Often, decision-makers choose one or a small range of preferred damage estimates, which suit their purposes or political preferences (e.g., Democrat vs. Republican administrations’ choices of the SCC), and ignore others. We provide specific examples of such practices in the Section 1. Our analysis suggests that these practices are misguided, and they risk producing misleading or unwise decisions.

They conclude:

The climate–economy relationship therefore remains deeply uncertain. Recognizing this uncertainty is not a failure of economics, but a necessary step toward more honest analysis and more robust policy and practice. Future research should focus less on extracting ever more precise estimates from insufficient data, and more on understanding the mechanisms of adaptation, resilience, and institutional change that will shape economic outcomes in a warming world.

Amen.

Important: If you appreciate being kept up-to-date on the latest climate research, please let me know by clicking that “❤️ Like.” More likes means more readers of THB. Thanks!

Comments welcomed!

Leave a comment

It is simple: THB exists because of its paid subscribers. Please consider joining them.

Subscribe now

Technical details below.

SOURCE NOTES for TABLE 1 — CB26 Replication Findings

Burke, Hsiang & Miguel (2015), Nature 527, 235–239 — Claims unmitigated warming cuts global GDP per capita 23% by 2100 via an inverted-U damage function peaking at 13°C. CB26 find that removing 6 influential outlier observations attenuates effect sizes by more than 20% and weakens the quadratic relationship, though significance does not disappear entirely. The effect does vanish in the post-2000 sample — a temporal instability with no agreed explanation. Newell, Prest & Sexton (2021) found no statistically significant temperature effect on GDP across 800 model specifications using BHM’s own data. CB26’s conclusion: the magnitude of any climate-economy effect cannot be reliably estimated from this methodology — not that no effect exists.

Kahn, Mohaddes, Ng, Pesaran, Raissi & Yang (2021), Energy Economics 104 — Claims climate damages accumulate permanently into GDP growth rates, with 3°C of warming cutting global GDP by up to 18%. CB26 find that removing 9 influential observations attenuates but does not eliminate the effect. More consequentially, KMN calculate projected damages from the acceleration of climate change rather than its level — a specification choice that produces the counterintuitive result of positive projected effects under low-emissions RCP2.6 for sub-Saharan Africa. Effects prove unstable across moving time windows.

Bilal & Känzig (2026), Quarterly Journal of Economics — Claims 1°C of warming reduces world GDP by more than 20% and puts the Social Cost of Carbon above $1,200 per tonne. CB26 find BK26’s results sensitive to ENSO variation, a bandstop filter, decade fixed effects, and moving time windows in the PWT dataset underlying BK26’s main damage function. Note: BK26 themselves showed their country-level local temperature results carry no significance at 5% — CB26 did not need to demonstrate this. CB26’s core objection is that BK26’s exclusion restriction — that short-run global temperature shocks carry the same indirect effects as long-run temperature changes — is difficult to justify.

Source: Curtin & Burgess (2026), “The empirically inscrutable climate-economy relationship,” SocArXiv. CB26 identify specification fragility and deep structural uncertainty — they do not call for retraction of any paper. The Kotz et al. retraction followed a data error, which is qualitatively different from the specification sensitivities CB26 document.

SOURCE NOTES for TABLE 2 — Where the Top-Down Climate-Economics Literature Shows Up in Policy and Finance

U.S. IWG Social Cost of Carbon — $51/tonne interim value (2021) | DICE model augmented with BHM-era damage functions from the Howard & Sterner (2017) meta-regression of top-down panel GDP studies. | Set the regulatory carbon cost used across the entire U.S. federal government for fuel economy standards, EPA rules, and infrastructure cost-benefit analyses.

EPA Proposed SCC — $190/tonne (2022) | BHM and updated damage functions from the RFF/GIVE model, which explicitly incorporates BHM-era panel regression estimates. | Proposed near-quadrupling of the U.S. regulatory carbon price; drove EPA power plant and vehicle rules; contested in federal court.

CBO Climate-GDP Analysis (February 2025) | Meta-analysis incorporating BHM, Burke & Tanutama (2019), Kalkuhl & Wenz (2020), Kotz et al. (2024, since retracted), and Nath et al. (2024). | Official federal projections of macroeconomic climate damages; CBO noted that Kotz drove substantial upward pressure on its damage distribution.

Biden White House CEA/OMB Climate Macro White Paper (April 2024) | BHM, Dell, Jones & Olken (2012), Kahn et al. (2021), Acevedo et al. (2020), Casey et al. (2023) — all top-down panel GDP studies. | Official executive-branch assessment of climate macroeconomic impacts; used to justify regulatory ambition and characterize risks from inaction.

NGFS Central Bank Scenarios — Phase V (November 2024) | Kotz et al. (2024, retracted) embedded as the baseline damage function; NGFS warned users to “be aware” of the retraction only post-publication. See Roger Pielke Jr.’s analysis. | Adopted by 130+ central banks — Bank of England, ECB, Federal Reserve — as the standard framework for climate stress testing and capital adequacy assessment.

Bank of England Climate Biennial Exploratory Scenario (2021–2022) | NGFS scenarios drawing on BHM-lineage damage functions; Kotz methodology embedded in subsequent NGFS phases. | Directed UK banks to quantify climate-driven losses using top-down damage functions; informed PRA supervisory capital expectations.

SEC Climate Risk Disclosure Rules (2024) | Integrated assessment models incorporating BHM-era damage functions; NGFS and TCFD scenario frameworks tracing to the same top-down literature. | Required ~7,000 public companies to disclose climate financial risks using top-down-derived methodologies; partially stayed by courts.

TCFD Corporate Climate Reporting Framework | Top-down GDP damage literature underpins the “physical risk” scenario quantification in TCFD guidance; BHM-lineage results travel through NGFS into TCFD disclosures. | Adopted globally; mandatory in the UK, EU, Canada, and Australia; trillions in assets under management now subject to TCFD-aligned climate reporting.

IMF Fiscal and Financial Stability Climate Risk Analyses | Kahn et al. (2021) long-run growth effects; BHM; related top-down panel literature informing IMF staff country assessments. | IMF estimates of sovereign debt sustainability, fiscal space, and macroeconomic stability draw on the top-down framework; shape IMF surveillance and lending conditions.

Climate Litigation — Expert Testimony & Damages Claims | BHM, KMN, Bilal & Känzig; high-end SCC estimates derived from top-down damage functions; used in Rhode Island v. Chevron, Hawaii, and other cases. | Expert witnesses argue fossil fuel liability in the trillions; BK’s $1,200/tonne SCC amplifies claimed damages by orders of magnitude.

Subscribe now

Today, I discuss a new pre-print by Finbar Curtin and Matt Burgess of the University of Wyoming which is by far the most important climate paper I’ve read in quite some time — “The empirically inscrutable climate-economy relationship.”1

Curtin-Burgess (CB26) ask a straightforward question: Can we actually measure how climate affects the economy from the historical record?

Their answer is no.

Economists have sought to identify a relationship of climate and the economy from the historical record as a basis for projecting into the future how changes in climate might affect economic growth. CB26 explain why meaningfully identifying that relationship is simply not possible.

In today’s post, part one of two, I explain the analysis of CB26 and take it further. I argue that the question they ask was never really answerable to begin with — the entire project of regressing aggregate economic output on aggregate climate variables is conceptually and fatally flawed.

This post briefly overviews the macroeconomics literature that has sought to connect climate and the economy, describes the theoretical framing of CB26, explains how I take their arguments even further, and sets the stage for Part 2 — which looks at CB26 replications of key climate-economics papers.

Grab a cuppa and settle in . . .

Bottom Up and Top Down

Two parallel traditions in climate economics ask two different questions.

The older tradition, which dates to William Nordhaus’s 1991 paper “To slow or not to slow” and the DICE model he introduced in 1992 asked: what is the right global carbon price?

Answers to that question estimate what has come to be known as the social cost of carbon — the dollar value of the damage of an additional emitted ton of CO₂.

This work typically follows a bottom-up approach: Nordhaus and those who followed him estimate climate damage sector by sector, such as agriculture, human health, energy demand and so on. Add up the sectoral damages, and get a total damage estimate.

Nordhaus won the Nobel Prize in 2018 for this work, and DICE remains the reference integrated assessment model for U.S. government analyses.

The bottom-up approach combines — sector-by-sector — various damage functions that relate temperature (typically) to economic outcomes. Sectoral damage functions make explicit the causal pathway from climate to economic impact.

The newer tradition — top-down — emerged in the 2010s and asks a related but different question: how does climate actually affect economic output in the historical record as a basis for projecting how changes in climate might affect the economy in the future?

The top-down approach seeks to bypass creating sectoral damage functions and instead looks to establish a relationship between climate variables (typically temperature) and aggregate economic output. The core idea is that once that relationship is established, it can be used to explore how future economic output may change based on changes in climate variables.

Dell, Jones and Olken (2012) is an early example of a panel regression of GDP growth on temperature. Burke, Hsiang and Miguel (2015) took this approach forward and produced the influential claim that unmitigated warming would cut global GDP per capita by 23 percent by 2100.

Many papers followed — Kahn et al. 2021, Kalkuhl and Wenz 2020, Kotz et al. 2024 (since retracted), Nath, Ramey and Klenow 2024, Bilal and Känzig 2026 — each relating an aggregate climate index to an aggregate economic index and interpreting the resulting coefficient as a causal damage function, which could then be used to project future damage as a function of future changes in climate.

Curtin and Burgess challenge whether any such top-down approach can answer the core question it asks and conclude that top-down approaches are fundamentally incapable of answering that core question.

CB26 take on an entire literature.

The Curtin and Burgess Critique

Curtin and Burgess ask what it would take for an aggregate top-down regression to actually recover a meaningful causal relationship, and they explore whether econometric methods allow for meaningful results.

CB26 argue that they do not: The climate-economy relationship, as measured from aggregate historical data, is — in their word — inscrutable.

To be clear, CB26 explicitly accept that human activity changes the climate, posing risks. In fact, their critique has nothing to do with the physics of climate science. Their critique is about economics, and specifically methods of econometrics.

CB26 show — rigorously and comprehensively — that the magnitude of any climate effects on economic growth, resulting from the top-down approach, cannot come out of the historical record with any confidence, no matter how sophisticated the econometrics.

The CB26 theoretical argument is straightforward but technically involved. To illustrate, I will use panel-data regression of country GDP on country temperature as a running example, because that is the approach most of the top-down literature uses. The general conclusion applies to any top-down approach.

The data are a panel — a table with one row per country per year. For instance, Rwanda 1994 is a row. India 1980 is a row. The United States 2010 is a row. Each row has a temperature, a precipitation, a GDP growth rate, and, depending on the approach, a handful of other variables. For example, 170 countries across 50 years gives 6,500 rows (170 x 50). That panel has two dimensions, country and year.

Climate economists want to know how a change in temperature causes a change in GDP growth. The methodological challenge is that temperature correlates with nearly everything. Countries with different temperatures have different institutions, religions, colonial histories, natural resources, neighbors, etc. Similarly, years with different temperatures have different oil prices, wars, technology, financial crises, etc. A naive regression of GDP on temperature risks mistaking spurious influences for a temperature effect.

Economists handle this challenge with a tool called fixed effects.

The Princeton University Research Guide on panel data explains:

“The fixed effects model assumes that the omitted effects of the model can be arbitrarily correlated with the included variables. . . . Fixed effects explore the relationship between predictor and outcome variables within an entity (country, person, company, etc.).”

In plain language: fixed effects tell the regression to throw away every kind of variation that is either specific to a country (and therefore potentially confounded with the country’s average climate) or specific to a year (and therefore potentially confounded with global shocks that hit every country the same way).

Country fixed effects remove everything specific to a country that stays constant over time — culture, institutions, geography, history. Year fixed effects remove everything that hit every country similarly in a given year — the 2008 financial crisis, the 1973 oil shock, global technology trends. What remains is a residual: India ran hotter than usual in 1987, cooler than usual in 1988 — and the regression uses that residual to identify the temperature effect.

This approach works only if the temperature-GDP relationship is the same everywhere, at every time. The regression pools every country’s residual and fits a single function — or a single quadratic, in the Burke-Hsiang-Miguel case — through the data.

CB26 explain:

“Many climate-econometric studies assume one global climate-economy relationship. They employ panel methods which treat every individual time-period observation as variation over this global function (e.g., Dell et al. (2012); Burke et al. (2015c); Kahn et al. (2021)).”

The relationship cannot be the same everywhere. Take two countries that look similar on key dimensions. CB26 point to El Salvador and Iraq: about the same average temperature and the same level of affluence. When the authors estimate a temperature effect for each country separately, they find a positive relationship for El Salvador and a negative one for Iraq. The average conceals a tropical climate (El Salvador) and a desert oil economy (Iraq), a remittance-dependent labor market and a war-scarred institutional environment.

Pooling El Salvador and Iraq into one regression is, according to CB26, methodologically unsound:

“Assuming a global climate-economy relationship means that points from one country implicitly affect the estimate of the climate-economy relationship for all countries. Fixed-effects models result in countries’ relationships being implicitly weighted by their variations in temperature (and/or other climate variables used), with larger-variance countries given higher weight. . . . The result is that relatively small countries at temperature extremes can have a large influence on global estimates.”

Not only is it problematic to pool relationships across space, it is also problematic to pool them across time. For instance, a climate-economy relationship in India in 1965 is not the relationship in India in 2015. Over that time India’s GDP increased by a factor of ten and saw expansion of air conditioning, irrigation, and a service sector.

Economic development changes how a country experiences weather: Kahn (2005) and Toya & Skidmore (2007) each found that a one-percent increase in real GDP per capita cuts death and damage rates from natural disasters by roughly half a percent. Rich countries and poor countries respond to identical weather differently, and the same country responds to identical weather differently as it grows rich. Pooling five decades of Indian data and looking for a temperature effect on GDP is, again, according to CB26, methodologically suspect.

Any relationship of climate and the economy varies across both of the panel’s dimensions and CB26 argue that assuming otherwise, as the top-down literature does, introduces bias.

An obvious fix would be to let the relationship vary — give every country its own coefficient, give every year its own coefficient, and let the coefficient depend on income and baseline climate. But that approach is also methodologically problematic because it turns a search for a fixed relationship into a need for as many relationships as there are specific times and places.

CB26 explain:

“Each data point in a climate-economy time series has a time index (t) and a space index (i, often country, sometimes region). If the climate-economy relationship has qualitatively meaningful variation across both space and time — as we argue it does — there are not enough degrees of freedom to estimate it. Assuming away some of spatial and/or temporal variation to preserve degrees of freedom creates qualitatively meaningful bias. This is the core, irreducible estimation challenge.”

So the top-down researcher has no way out. Let the climate-economy relationship vary as much as it actually varies, and it cannot be modeled. Hold it fixed, and the model can generate a headline-producing result, but that result has no real-world meaning in any specific place.

Curtin and Burgess describe this as the ”core, irreducible estimation challenge” that better data or cleverer methods cannot fix.

CB26 go to some length to explain that theirs is a statistical argument, not a climate argument. It is a problem any panel-data researcher faces when the when the relationship of interest varies along every dimension of the data. Climate econometrics sits in an especially bad spot because both dimensions of heterogeneity are large.

The methodological problem that Curtin and Burgess describe is well-established in modern econometrics — they cite Chernozhukov et al. 2013 from that literature.

What is new in the Curtin-Burgess paper is the argument that this weighting problem is not simply a minor methodological issue — it is a problem basic to the structure of every pooled estimate in the field.

Spatial and temporal heterogeneity are the core of the identification failure, but they are not the only problems, and Curtin and Burgess discuss three others.

• The first is the problem of growth miracles and growth disasters, terms CB26 use to describe a small number of extreme economic events in the historical record that have nothing to do with climate — such as the collapse of the Soviet Union, the Rwandan genocide, Iraq’s post-war rebound after 2003, or Oman’s 1968 oil boom. Such events produce single-year GDP changes ten-to-thirty-times larger than a country’s typical growth rate, up or down. When one of them happens to fall in a year of unusual weather in a small country, the regression treats the weather as the cause of the GDP swing. Curtin and Burgess show that six or nine such observations — out of roughly 6,500 in the dataset — drive a quarter to a third of the estimated climate damage in prominent papers (and will be discussed in Part 2).

• The second is that the climate a country has already adapted to is not observable in the data. An economist cannot distinuish damage from novel events from damage caused by expected events in a country being poorly matched to its environment. A cold country with poor winter infrastructure will suffer losses when winter happens, but calling that climate damage would obviously be misleading, because the cause is poor adaptation to documented climate variability.

• The third is that “climate” is not one number. Temperature, precipitation, humidity, extreme events, ocean circulation, and wind patterns all matter for how society functions and these variables interact with each other. Nearly every top-down study collapses all of that into annual-average temperature and total precipitation.

Why I Think the Problems Go Even Deeper

Here go beyond Curtin and Burgess — not to dispute their argument, which is solid, but to push it further than they do.

The authors treat the problems above as technical challenges in econometrics: Spatial heterogeneity introduces bias. Temporal heterogeneity introduces bias. Influential observations introduce fragility. Reducing climate to one variable introduces omitted-variable bias.

Their framing suggests that, in principle, a top-down regression done carefully enough with good enough data would recover a meaningful relationship if only researchers could overcome these obstacles. CB26 argue that in practice researchers cannot overcome the obstacles, so the relationship is “empirically inscrutable.”

Their theoretical model presumes that a well-defined climate-economy relationship exists and that the problem is aggregating it into a single estimable function based on available data.

I go further — the aggregate relationship, as the top-down literature conceptualizes it, simply does not exist as a coherent causal object at all.

Consider what the top-down regression actually does when seeking to relate an average temperature to GDP. Both are indices: No person, no crop, no factory, no piece of infrastructure ever experiences average temperature. A farm in Iowa experiences the temperature at that farm, minute by minute. Similarly, a factory in Bangalore experiences the temperature at that factory. An average temperature in these studies is built up by averaging gridded observations across a country’s land area — usually population-weighted — and then averaging again over the year. What enters the regression is a statistical summary of a summary.

GDP is a similar kind of index. Country GDP aggregates the output of millions of firms, each operating under local conditions, across sectors whose relationships to weather differ enormously. Global GDP aggregates country GDPs. No person, company, or government experiences GDP.

A defender of the top-down approach might reply that GDP is the variable we actually care about for policy, so treating it as downstream of the real mechanisms misses the point. That defense fails for a simple reason. GDP is the outcome we care about, and climate change is the cause we care about, but neither aggregate index is where the causal process actually plays out. The question is what causes changes in GDP — and those causes operate at the level where people, firms, and physical systems respond to their local conditions.

Let me illustrate the problem with an analogy. Imagine trying to learn whether paying baseball players more makes them hit better by comparing team payrolls to team batting averages. You would find a correlation — higher-payroll teams do tend to hit better on average, because they pay for talent — but the correlation tells you almost nothing about the underlying question about how pay influences a batter’s hitting.

A team’s payroll is the sum of what every player on the roster earns, spread across superstars and journeymen in different proportions on every team. A team’s batting average is a blend of what every hitter contributed, spread across starters and bench players in different proportions on every team across a long season. Whatever mechanism links pay to performance operates at the level of individual players in individual at-bats, and neither aggregate metric provides the data necessary to establish a causal relationship.

Correlating the two team averages can give a number, but it is a statistical artifact of thirty different roster structures over 162 games per team, with forty or so different players suiting up over a season, playing in different stadiums against different opponents.

Fancy econometrics can certainly produce a number that relates an index of batting to an index of salaries. The number will reflect some real pattern — higher-paying teams do get better hitters on average — but it will not answer the question a team owner actually cares about, which is whether spending more money on this roster or on new players would improve this team’s performance. That counterfactual lives at the level of individual players, individual contracts, and individual at-bats, not at the level of team or season aggregates.

The climate-economy regression is worse off, because the indices it correlates sit even farther from the underlying mechanisms than payroll and batting average do.

The bottom-up approach, for all its problems, at least tries to include plausible causal pathways between changes in weather variables and outcomes. You can look at a Nordhaus damage function and see exactly what the author thinks is happening in agriculture, on coastlines, in the energy sector as climate changes.

The top-down approach skips the causal mechanism step and assumes that the aggregate statistical relationship somehow reflects the sum of all causal mechanisms.

CB26 provide strong evidence that the assumption fails. The sensitivities they document across the top-down literature — 20, 30, 50 percent swings in headline estimates from minor specification changes; wildly different answers from papers using the same data; results driven by a handful of observations that have nothing to do with the weather — are exactly what we should expect from a regression that is correlating two indices with no stable causal pathway between them. If a real aggregate relationship existed and had a stable causal interpretation, it would show up consistently across specifications that reasonably isolate it. The top-down climate literature’s estimates do not show any such consistency.

The field has mistaken the inscrutability of a statistical artifact — a quantitative relationship between two abstract indexes — for uncertainty about a presumed-real quantity. The damage function the top-down literature claims to estimate simply does not exist as a coherent object.

If the problem were merely technical — better data, longer panels, more clever econometrics — the field could reasonably expect to close in on the truth over time. If the problem is conceptual — that the top-down approach is correlating two downstream summaries with no causal structure between them — then more data and cleverer methods will not help. The field will continue producing numbers and those numbers will generate headlines, but those numbers, whatever they mean, don’t tell us anything meaningful about how future changes in weather might affect future economic activity.

In Part 2, I walk through what Curtin and Burgess found when they reproduced the three most influential papers in the top-down literature — and what they document is devastating.

Read their pre-print here.

Comments welcomed!

Leave a comment

Before you go: If you appreciate being kept up-to-date on the latest climate research, please let me know by clicking that “❤️ Like.” More likes means more readers of THB. Thanks!

THB exists because of its paid subscribers. Please consider an upgrade!

Subscribe now

Subscribe now

Through yesterday, according to preliminary data from the National Oceanic and Atmospheric Administration (NOAA) Storm Prediction Center (SPC), the U.S. has experienced 365 tornadoes, just four above the longer-term average to date. Last year, more than 1,900 tornadoes were observed, the most since 2011, and well above the longer-term average.

Today, I share the latest data on normalized U.S. tornado losses since 1954 and a time series of the incidence of the strongest tornadoes since 1975. I doubt you’ll come across these data anywhere else.

Comparing tornado losses across decades is not straightforward. A tornado striking what was a rural county 70 years ago would cause far less damage than the same tornado striking that same heavily developed county today — because there is now more property and wealth exposed to loss.

To make meaningful historical comparisons of loss estimates, researchers “normalize” losses: they ask what each historical storm or event would cost if it occurred under today’s societal conditions, adjusting for factors such as inflation, wealth, building types and counts, population, and, in some cases, efforts to improve building quality.

This approach was first applied systematically to tornadoes by Simmons, Sutter & Pielke Jr. (2013), which analyzed NOAA SPC data from 1950 through 2011. Zhang et al. (2023) reproduced and extended the Simmons series through 2018, confirming the earlier results and updating the time series. They concluded: “[O]ur results suggest a downward trend in tornado losses for the U.S. as a nation.”

The figure below shows my replication of the Zhang et al. tornado normalization from primary data and extends it through 2025. The normalized losses are expressed in 2026 dollars.

The dominant loss years are 1954 ($36 billion),1965 ($44 billion), and 1974 ($29 billion). The largest recent loss year is 2011 at $16 billion — the largest post-1980 value and the only recent year approaching the scale of the 1960s–1970s peaks. Since 2012, annual normalized losses have largely remained below $5 billion.

The time series shows a significant decrease in annual normalized losses.The 1954–1963 decade averaged $4.8 billion per year; the 2015–2025 decade averaged $1.9 billion per year.

Our 2013 paper identified this trend:

“We can definitively state that there is no evidence of increasing normalized tornado damage or incidence on climatic time scales.”

At the time, we hypothesized that underlying the decrease may be an actual reduction in severe tornado incidence. However, because economic data should not be used to infer trends in related climate variables, we suggested that any such trend in tornadoes would depend upon analyses of climate data.

More than a decade later, tornado data is strongly suggestive of an overall decline in the incidence of the strongest tornadoes. The figure below shows the annual count of F3/EF3 and stronger tornadoes from 1975 through 2024, the period over which data quality is most consistent.

The time series shows a clear decrease in major tornado incidence. The 1975–1984 decade averaged 49 F/EF3+ tornadoes per year; the 2015–2024 decade averaged 26 per year — a decline of roughly 46 percent. Note that data available back to 1954 makes this decrease look much larger, but is accompanied by questions of data quality.

The interpretation of this declining trend requires caution for several reasons. First, the Enhanced Fujita (EF) scale introduced in 2007 changed rating methodology, creating a potential discontinuity in the series. Most analysts believe EF ratings are somewhat more conservative than legacy F ratings for comparable damage, which could contribute to the apparent post-2007 decline. Second, improved public warnings and storm-resistant construction may have contributed to changing the nature of observable damage markers that drive intensity ratings, which are often established based on damage patterns rather than direct measurements of tornado intensity.

Neither of these caveats undermines the the fact that there is no evidence of an increase in violent tornado incidence over the observational record. The IPCC Sixth Assessment Report concluded with low confidence in the detection of any trend in tornado frequency or intensity at the global or regional level, and with low confidence in attribution of any observed changes to anthropogenic forcing. The data reviewed here are consistent with that assessment.

Several hypotheses have been offered in the literature for how accumulating greenhouse gases in the atmosphere may influence tornado behavior. Some commonly cited examples include:

• Reduced wind shear. Trapp & Hoogewind (2018) proposed that Arctic amplification could weaken lower-tropospheric wind shear, reducing the environmental favorability for supercell formation. If correct, this would represent a climate signal leading to fewer strong tornadoes.

• Geographic shift eastward. Gensini & Brooks (2018) documented a westward decline and eastward increase in tornado activity — a shift toward “Dixie Alley” and away from the traditional “Tornado Alley.” An eastern shift moves tornado tracks closer to higher-density development. They conclude: “At this point, it is unclear whether the observed trends in tornado environment and report frequency are due to natural variability or being altered by anthropogenic forcing on the climate system.”

• Increased variability. Brooks, Carbin & Marsh (2014) argued that over 1954 to 2013 tornado activity became more variable — with years of very high activity alternating with years of historically low counts — rather than showing a simple trend in frequency. They explain: “At this point, we cannot offer a physical hypothesis for the increased variability.”

The normalized loss series and the F/EF3+ incidence series both show the same pattern: no upward trend, and arguably, a significant downward trend that has contributed to lower normalized damage levels in recent decades.

Before you go: If you appreciate being kept up-to-date on the latest data on extreme weather, please let me know by clicking that “❤️ Like.” More likes means more readers of THB. Thanks!

Comments welcomed!

Leave a comment

Note: For paid subscribers, a full replication spreadsheet with all annual data, normalization factors, and source documentation is available for download at the bottom of this post.

For Further Reading

Ashley, W. S., Haberlie, A. M., & Gensini, V. A. (2023). The future of supercells in the United States. Bulletin of the American Meteorological Society, 104(1), E83–E105. https://doi.org/10.1175/BAMS-D-22-0126.1

Brooks, H. E., Carbin, G. W., & Marsh, P. T. (2014). Increased variability of tornado occurrence in the United States. Science, 346(6207), 349–352.

Gensini, V. A., & Brooks, H. E. (2018). Spatial trends in United States tornado frequency. npj Climate and Atmospheric Science, 1, 38.

Simmons, K. M., Sutter, D., & Pielke, R. Jr. (2013). Normalized tornado damage in the United States: 1950–2011. Environmental Hazards, 12(2), 132–147.

Trapp, R. J., & Hoogewind, K. A. (2018). Exploring a possible connection between U.S. tornado activity and Arctic sea ice. Journal of Climate, 31(2), 571–587.

Verbout, S. M., Brooks, H. E., Leslie, L. M., & Schultz, D. M. (2006). Evolution of the U.S. tornado database: 1954–2003. Weather and Forecasting, 21(1), 86–93.

Zhang, Y., et al. (2023). Time trends in losses from major tornadoes in the United States. Weather and Climate Extremes, 41, 100639.

Note: For paid subscribers, a full replication spreadsheet with all annual data, normalization factors, and source documentation is available for download below.

Read more

Subscribe now

This week, I am returning to the University of Colorado Boulder campus for the first time since I retired from the university in 2024. I was invited to participate in the university’s Conference on World Affairs, where I’ll participate on two panels tomorrow — both livestreamed: one on climate change (9:30AM MT) and one of federal funding for research (2:00PM MT). Today, I’ll guest lecture in an upper-level atmospheric science course (buckle up students!).

I appreciate the invitation, surely a positive sign of an intellectual thaw. At the same time, the university gave in to a demand that I be removed from the original climate panel I was assigned to — so they created a new panel for me to be on. But any progress is good progress. If you come to my panels tomorrow, please do say Hi!

Returning to campus prompted me to take a look back at my posts here at THB on problems in U.S. universities, in the context of recent data indicating that public confidence in U.S. higher education has rebounded a bit off the lows of recent years, as shown below.

Over the past few years I’ve had the good fortune to visit more than a dozen universities, to meet with university leaders in public and private institutions, and to participate in AEI’s new Center for the Future of the American University. The small rebound in public opinion reflects what I’ve observed firsthand: there are indeed some positive signs that university leaders are correcting course. There is much work still to do.

I have updated and integrated my writings on ‘the university problem’ into a report — the first in what will become a continuing THB feature: Deep Dive Reports. The idea: synthesize a series of arguments and data into a single accessible document — more in-depth than a single post, less sprawling than a long series. All reports will have updated data and figures that THB readers are free to use and share.

Deep Dive Reports are also an experiment in delivering more value to paid subscribers — please do let me know what you think.

Below is the abstract of the report, which paid subscribers can download in full below.

Abstract

Public confidence in American higher education has rebounded modestly in 2025 — the first measured increase in a decade — but the numbers remain far below where they stood when Gallup began tracking them in 2015, and the underlying drivers of skepticism have not resolved. This report draws on new polling from Gallup/Lumina and the Vanderbilt Unity Poll alongside a decade of survey data to diagnose what has gone wrong and what it would take to fix it.

The conventional framing — declining university confidence as a Republican problem — turns out to be badly incomplete. Trust in scientific and academic institutions has fallen steeply among Hispanic and Black Americans, among Democrats without college degrees, and among working-class voters of both parties. The one group reporting high and rising confidence in universities is the same demographic from which universities disproportionately recruit their faculty and administrators: highly educated, secular, wealthy, white liberals. That alignment is not a coincidence. It is the core of the problem.

The report documents the dramatic leftward shift in faculty political composition since the early 2000s and traces the institutional mechanisms — above all the NSF’s 1997 “broader impacts” criterion and the post-Bush-era alignment of science with partisan Democratic politics — that reflected a turn from individual faculty advocacy into institutionalized political action. It draws on the author’s nearly twenty-four years as a tenured professor at the University of Colorado Boulder to show what that institutionalization looks like on the ground. And it evaluates the current debate over institutional neutrality, arguing that what universities actually need is not neutrality — an impossibility — but institutional restraint: a strong, enforced presumption against the university as an institution taking sides in external political controversies.

Three steps toward recovery: return to the academic mission, take institutional restraint seriously in practice rather than just in policy documents, and expand access so that universities serve all Americans rather than the narrow demographic whose values they currently reflect.

Comments welcomed! I invite your recommendations for future Deep Dive Reports.

If you like the idea of occasional THB Deep Dive Reports —> Please let me know by clicking that “❤️ Like.” More likes means more readers of THB. Thanks!

Read more

Subscribe now

Last month I documented how the Reference Manual on Scientific Evidence (4th ed., 2025)1 included a chapter on climate change that was written by authors whose work supports climate litigation, and, even worse, was ghost written by a plaintiff’s attorney in climate cases. The Manual — published by the Federal Judicial Center (FJC) in partnership with National Academies of Science, Engineering, and Medicine (NASEM) — is supposed to offer a balanced treatment of the myriad complexities where science meets the legal system, as a reference for federal judges.

Today, I return to the FJC Manual to discuss another of its chapters: How Science Works, which underlies the entire volume. The chapter appears to be telling judges “how science works” in a manner supportive of the climate chapter’s characterization of climate litigation.

In her critique of the How Science Works chapter, Jessica Weinkle takes issue with how the chapter presents scientific consensus as the basis of policy action and legal judgment, a form of the “linear model” of science and policy:

By making consensus reporting a necessary precursor to action and an inevitable outcome of scientific research, the linear model circumscribes messy but legitimate democratic decision-making processes. Chapter 2 literally waves away the “twists and turns” that enabled success and, in so doing, positions itself as the world champion of scientific consensus reports- like those produced by NASEM. The implication is that decision-makers need only fall in line with the reasoning and worldviews of scientists.

The fable is false, cliché-ridden, and, in today’s age, utterly tone-deaf. Nonetheless, the FJC-NASEM powers that be saw it fit to feature it in the latest Reference Manual.

Throughout, the How Science Works chapter uses the experience of the 1974 discovery that chlorofluorocarbons (CFCs) deplete the ozone layer and its aftermath to illustrate how science “worked exactly as it ought” in policy and legal settings.

The chapter argues that in the case of ozone depletion, scientific research marched inexorably toward consensus, overcame industry-supported “merchants of doubt,” consensus motivated policy action, and crucially, informed legal action based on science that was beyond question.

In reality, all of this is wrong.2 This post explains why.

Let’s begin my detailed critique of the retelling of the ozone story in How Science Works, with a long excerpt from the chapter:

[Mario] Molina and [Sherwood] Rowland published an article [in 1974] putting forth the hypothesis that CFCs deplete our protective ozone layer, but their evidence did not come from an experiment or even any new observations. Instead, the pair brought together measurements collected by other researchers, calculations, and established knowledge about basic and atmospheric chemistry, arguing that if all these other ideas and estimates were true, a logical outcome is that CFCs pose a threat to the ozone layer. Later, experiments were performed that suggested the chemical reactions that Rowland and Molina reasoned should happen actually did happen. Other scientists incorporated these ideas into their models of the atmosphere and made predictions about what should be observed if Rowland and Molina’s hypothesis were correct. Still other groups collected atmospheric evidence to test the predictions made by the models. Meanwhile, the CFC industry backed another scientist to oppose the hypothesis, and Rowland and Molina checked some of the old measurements on which they had based their hypothesis and found them to be inaccurate. After correcting those numbers, models were updated, compared, and updated again. More data were collected and eventually, eight years after the hypothesis was first published, researchers discovered a thinning of the ozone layer over Antarctica much more extreme than expected. This led to more revisions of the hypothesis, a few additional twists and turns, and eventually a Nobel Prize, the Montreal Protocol, which phased out CFC production, and today, a recovering ozone layer. Concurrently, ozone depletion has made its way into the courts as established science. For example, the National Resource Defense Council (NRDC) brought suit against the Environmental Protection Agency (EPA) for making decisions that violated the Montreal Protocol. The NRDC was found to have standing because of the increased risk of skin cancer that NRDC members would experience as a result of ozone destruction. Ozone depletion was treated as a fact in that case, reflecting the scientific consensus on the issue.

There are at least four major problems with this retelling of the ozone depletion experience.

Problem 1: Policy Action Occurred as Science Became Less Certain and More Contested

The FJC Manual tells a story of steadily accumulating scientific evidence converging on consensus, despite a few “twists and turns.” In the decade after Molina and Rowland’s seminal 1974 paper, authoritatve scientific assessments moved away from a widely-shared consensus on the threat of ozone depletion.

The chapter states:

“While the complex and iterative processes that went into establishing depletion of the ozone layer by CFCs are commonplace in science, the speed with which societal and political action followed scientific consensus in this case may be unusual.”

The real lesson is that political action came well before consensus.

A 1976 NAS (National Academy of Sciences) report confirmed the Molina-Rowland finding and projected 15–18% ozone depletion over a century — lending support to the 1978 U.S. aerosol ban.

But successive NAS assessments revised that estimate steadily downward:

• A 1979 NAS report projected 11–16% century-scale decline;

• A 1982 NAS assessment reduced that to 5–9%;

• And a 1983 NAS update projected just 2–4%.

These assessments reflected legitimate views of many leading scientists and they were invoked by those wanting to slow down continued tightening of regulations. As the Christian Science Monitor noted in March 1984:

“This successive downgrading of what some scientists had called a significant danger suggests that the United States acted more out of fear than foresight in banning the suspect chemicals. . . To begin with, the issue is not at all clear-cut scientifically. . . At no point has any expert claimed to have substantial proof that CFCs do, in fact, threaten the ozone layer. Even those experts most strongly in favor of banning the chemicals have urged this as a matter of prudence, not as a response to a clearly established hazard.”

The real lesson here — one not shared with federal judges — is that significant policy action took place from 1974 to 1984 despite the uncertainties and lack of a strong scientific consensus on the risks of significant ozone depletion.

Problem 2: The FJC Chapter Rearranges the Timeline

The FJC Manual works hard to spin a tale of policy driven by consensus and, to do so, mistakenly recounts the historical timeline. The chapter argues, incorrectly,

“eight years after the [Molina and Rowland] hypothesis was first published, researchers discovered a thinning of the ozone layer over Antarctica.”

Molina and Rowland published in June 1974. Eight years later is 1982. But the Farman, Gardiner, and Shanklin Nature paper documenting the thinning of the ozone layer over Antarctica — popularized as the “ozone hole” — was in 1985, eleven years after Molina and Rowland.

The three-year discrepancy matters enormously to the Chapter’s story because of what happened in those three years.

The Vienna Convention for the Protection of the Ozone Layer — the foundational international agreement that set the stage for the Montreal Protocol — was signed on March 22, 1985 —and negotiated during an extended period of scientific debate and uncertainty, and crucially, before the discovery of the ozone hole was announced or popularized.

The real lesson here is that meaningful policy action took place in the face of considerable scientific disagreement over the magnitude and timing of the effects of CFCs on the ozone layer.

The 1985 ozone hole paper may have added momentum to the adoption of the 1987 Montreal Protocol under the Vienna Convention — Though it is worth noting that the U.S. EPA explained when proposing a rule to comply with the Montreal Protocol that, “the Agency has de facto assumed that the ozone hole is not related to CFCs and halons.”

The accurate lesson to take from the ozone experience is that substantial policy action took place in the United States and internationally from 1975 to 1985, a decade during which there was little consensus on the observed and projected effects of CFCs on the ozone layer.

The FJC manual is a bit too clever in its summary the timeline of events, explaining that policy success start with research that led to

“more revisions of the [1974] hypothesis, a few additional twists and turns, and eventually a Nobel Prize, the Montreal Protocol, which phased out CFC production, and today, a recovering ozone layer.”

Again, the chapter plays fast and loose with the timeline. The Montreal Protocol, which initiated negotiations toward a full phase out of CFCs, was agreed to in 1987. Molina and Rowland were awarded the Nobel Prize in Chemistry in 1995. The chapter implies the opposite.

The recovery of the ozone layer is underway, but detection of the effects of phasing out CFCs is slow and uneven, and detection of recovery in many places beyond Antarctica is not yet expected. On the 30^th anniversary of the Montreal Protocol, a group of atmospheric chemists asked, Is global ozone recovering? and they answered:

“The clearest signs of increasing ozone, so far, are seen in the upper stratosphere and for total ozone columns above Antarctica in spring. These two regions had also seen the largest ozone depletions in the past. Total column ozone at most latitudes, however, does not show clear increases yet. This is not unexpected, because the removal of chlorine and bromine from the stratosphere is three to four times slower than their previous increase. Detecting significant increases in total column ozone, therefore, will require much more time than the detection of its previous decline.”

The nuance matters. In the ozone case, policy action needed to occur well before the effects of policy implementation could possibly be detected, which would have to wait many decades. That means that agreement on policy is far more likely when actions have multiple justifications, and not just science.

Problem 3: Industry’s Role Was Actually Essential to Political Action

The FJC retelling of the ozone experience suggests simplistic industry opposition to the findings of Molina and Rowland in 1974: “the CFC industry backed another scientist to oppose the hypothesis.”

The scientist is Richard Scorer, who was an accomplished atmospheric scientist and professor at Imperial College London. He participated in a U.S. speaking tour in 1975, sponsored by a chemical industry PR group, calling the ozone depletion hypothesis “pompous claptrap.”

The FJC chapter uses Scorer to suggest how judges can evaluate experts based on their supposed positioning with respect to a consensus. Scientists who disagree with a consensus are likely to be industry shills. Not mentioned was the fact that Scorer’s industry-backed tour occured in the months after the original ozone depletion hypothesis was put forward, long before there was anything resembling a consensus.

The chapter explicitly generalizes its argument to “issues such as climate change and tobacco exposure.” Industry is reduced to a cartoonish roll of supporting “merchants of doubt,” a concept cited explicitly in the chapter, and connected to tobacco and climate change.

The chapter’s account of the role of industry in the ozone case is limited to the 1975 PR campaign, intimating that industry was opposed to policy action throughout. That is highly misleading.

DuPont — the world’s leading CFC producer with roughly 25% global market share — had identified HFC-134a as a Freon replacement by 1976 and patented the manufacturing process by 1980. By the time of the Montreal Protocol negotiations, DuPont was already interested in alternatives because their original CFC patents were due to expire. DuPont had filed more than 20 patents for CFC alternatives and its Freon division head explained the company’s interests: “There is an opportunity for a billion-pound market out there.” DuPont announced in 1988 that it would phase out CFC production entirely.

One reason why agreement was reached on the Montreal Protocol was that the availability of technological substitutes for CFCs helped to align the interests of environmental groups, governments, and industry around a shared outcome. There is a lesson here as well — As Karen Litfin explains in Ozone Discourses:

“Without regulation there could be no substitutes but, at least in the minds of many, without the promise of substitutes there could be no regulation.”

The technological chicken and the regulatory egg offered a lesson worth sharing with the judges, as it emphasizes the role of technology as a key factor in connections of science and policy.

The 1975 Scorer episode highlighted in the FJC chapter was a sideshow and ultimately inconsequential. The DuPont patent strategy that began in 1976 was a crucial factor in how technological innovation helped to connect policy action with scientific understandings.

The chapter emphasizes a sideshow and ignores the lesson that industry often plays a central role in both science and science in policy. That role is of course not always positive, but it is not always negative either. Judges should have been exposed to these complexities in the Reference Manual.

Problem 4: The Retelling Spins a Lawsuit as A Positive Lesson

The FJC Manual’s retelling of the ozone story concludes with a supposed lesson for those using lawsuits based on science:

“[T]he National (sic)3 Resource Defense Council (NRDC) brought suit against the Environmental Protection Agency (EPA) for making decisions that violated the Montreal Protocol. The NRDC was found to have standing because of the increased risk of skin cancer that NRDC members would experience as a result of ozone destruction. Ozone depletion was treated as a fact in that case, reflecting the scientific consensus on the issue.”

This is wrong on multiple counts.

Most importantly: NRDC lost this case on its merits — specifically, the court rules that international agreements under the Montreal Protocol are not U.S. law — and this is not shared in the chapter. Science did not compel a verdict in favor of the plaintiffs. In fact, the case turned on legal matters and scientific consensus was not relevant to the outcome. None of this is shared in the FJC Manual, which instead tells judges, “ozone depletion has made its way into the courts as established science.”

The lesson of the NRDC vs. EPA case highlighted in the chapter is that science does not compel legal outcomes.

What a Broader Look at the STS/STP Literature Would Have Told Judges Instead

The revisionist history of ozone depletion suggests a tidy story:

The FJC Manual selectively invokes some of the literature of science and technology studies (STS), but neglects some of the most important traditions in that literature that would have offered a very different perspective for Federal judges.

For example, Silvio Funtowicz and Jerome Ravetz developed the concept of “post-normal science” specifically for situations where “facts are uncertain, values in dispute, stakes high and decisions urgent” — the conditions that often characterize science in litigatio. Funtowiz and Ravetz argue that in such situations ordinary markers of scientific quality are least reliable and the most important decisions are made not after consensus is achieved, but in the presence of irreducible uncertainty, conflict, and ignorance.

The ozone case is a textbook example of post-normal science. Meaningful action occurred well before consensus was established and, in fact, while authoritative risk estimates were declining. Explaining post-normal science would have helped judges to understand this reality rather than telling them that consensus leads action.

In another example, Daniel Sarewitz argues that more science does not necessarily resolve environmental controversies — it can exacerbate them by supplying contesting parties with their own bodies of legitimate facts, an “excess of objectivity.” The successive NAS revisions of the ozone depletion estimate are a perfect illustration: improving science produced declining risk estimates that opponents of regulation used to argue for inaction, while proponents argued that the remaining uncertainty justified precaution.

Sarewitz’s framework would equip judges to recognize this dynamic — and to understand that their role is not simply to identify which side has the scientific consensus, but to weigh evidence and make legal judgments even when the science is unsettled. In fact, uncertainty can underpin action, as happened in the ozone case.

This literature — and much more — are absent from the How Science Works chapter.

The issues in the How Science Works chapter are significant, and taken together with the problems of the climate chapter, suggest that the FJC and NASEM should be asked by Congress to explain how the corruption of the FJC Manual happened and what steps will be put in place to ensure that it does not happen again.

Before you do anything else —> Please click that “❤️ Like.” More likes means more readers of THB. Thanks!

Comments welcomed!

Leave a comment

THB exists because of its paid subscribers. Please consider upgrading and join in on the discussion. Thank you.

Subscribe now