THIS WEEK'S RMN
MARKETS · China Re Bets on EV Exports and Belt and Road to Drive Global Risk Market Share
China Re is also expanding use of DeepSeek AI in its risk workflow.. — Read →
MODELS · US Standing in Weather Modeling Draws Congressional Scrutiny
The Trump Administration’s NOAA chief fielded questions from lawmakers on AI modeling and a $100M-plus budget increase, but the agency's global ranking in weather forecasting hung over the testimony. — Read →
MARKETS · Data Centers Will Delay the AI Catastrophe Kill Switch Unless Operators Face the Cost
Data centers facing a rogue AI incident have rational financial incentives to delay pulling the internet kill switch, and the only thing likely to change that calculus is liability. — Read →
Iran’s Other Time Bomb
The Iran war is tied up with myriad risks that markets are struggling to measure and model, including energy, geopolitical and economic.
But in a risk markets increasingly battered by black swans, Iran and the countries involved in the current conflict face a compounding risk that could upend the region yet again: earthquake.
"I wouldn't say the next one is overdue," says Edwin Nissen, a professor of Earth and ocean sciences at the University of Victoria who has spent his career studying Iranian fault systems. "But I also wouldn't say the next one couldn't happen tomorrow."
Tehran alone has been destroyed by earthquakes four times in recorded history: 855 AD, 958, 1177, and 1830. With the last catastrophic quake striking the country 196 years ago, the scientific consensus on what the next earthquake would mean for a megacity of fourteen million people is not reassuring.
Ilan Kelman, a professor at University College London who studies the intersection of conflict and disaster, says that a worst-case scenario — a large quake in the near future — could produce the first recorded million-death earthquake in human history, in a region already battered by war. Nissen agrees. A scientist who has worked in Iran since before the 1979 revolution told him a large enough quake would produce a similar outcome. "In his opinion," Nissen says, "the first million fatality earthquake — the first earthquake to kill over a million people — could well be Tehran."
None of the geology behind those projections has changed since the US and Israel began operations against Iran on February 28. What has changed is everything the risk models assume around it.
The Same Forces That Made the Oil Drive The Quake Risk
To understand why Tehran's earthquake risk matters to the energy markets and the risk market simultaneously, it helps to understand that they are not separate stories. They are the same story, told by the same geology.
"Iran is kind of trapped within a collision between two big tectonic plates," Nissen explains. To the north is the Eurasian plate, encompassing everything from Western Europe across to eastern Russia and China. To the south is the Arabian plate, a smaller piece of crust that has been rifting off Africa over tens of millions of years. The Red Sea exists because of that separation — its opening is what pushes Arabia steadily northward.
"You can think of Iran as being like the front of a car in a car crash," Nissen says. That crash is also what made the Persian Gulf's oil.
The northern edge of the Arabian plate carries a sedimentary sequence roughly ten kilometers thick — layers of organic-rich source rock, fractured reservoir rock, and impermeable cap rock, everything needed to generate and trap hydrocarbons. What makes Iran exceptional, Nissen says, is that the collision is actively folding that cap rock into great underground arches. "The oil percolates upwards and gets trapped within these upfolds — we call them anticlines." The Zagros Mountains, which follow the Iran-Iraq border all the way down to the Strait of Hormuz, are the surface expression of those folds.
"The Zagros is the single best reservoir for oil and gas in the world, bar nowhere else," he says. "It's got the source, the reservoir, the cap, and it's got these folds."
The folds, however, are not the whole story. "The folds are also related to faults," Nissen says, "because the shallow part of the crust is being folded, but underneath these folds there are also faults — and that comes back to the car wreck." The same collision that traps the oil generates the earthquakes. They are not coincidental features of the Iranian landscape. They are products of the same force.
There is a further layer. If you look at a night-sky satellite image of Iran and overlay a fault map, Nissen says, the correlation is immediate: people live along faults. This is not accident. "The mountains are being uplifted along active faults, and those active faults are in exactly the same spot — at the edge of the mountain ranges. And the people also live along the edge of the mountains because that's where they get water."
Rain falls on the high ground, flows downhill, and surfaces as springs at the mountain front — which is also where the fault runs. "So there's this really interesting kind of fatal attraction between people and earthquakes in Iran." Tehran, at the foot of the Alborz range, is the largest expression of that fatal attraction on earth.
Why Buildings Are the Problem
Nissen is careful about one word that appears frequently in coverage of Tehran's seismic risk. "I would avoid using the word overdue," he says. "I think it's a loaded term." Iran has so many active fault systems — unlike California's singular focus on the San Andreas, or the Pacific Northwest's on the Cascadia megathrust — that framing any single fault as overdue misrepresents how the risk is distributed.
What he will say is that Tehran is "the biggest single worry in Iran," and the reason is not the geology. It is the buildings.
The 1978 earthquake in Tabas, a small desert town, killed 11,000 people from a population of 13,000 — nearly 90 percent. The 2003 Bam earthquake, magnitude 6.6, not a particularly large event, killed between 26,000 and 30,000 people, roughly half the city's population. "Earthquakes don't kill people," Nissen says. "Buildings do." In Iran, the buildings are predominantly unreinforced masonry or mud brick, which he notes "just crumbles" under strong ground motion. Following Bam, Iran adopted California's seismic building code, essentially translating it into Farsi. The problem, Nissen says, is enforcement. "It's such a corrupt country that you pay off the inspector and you don't need to follow the code. So even modern buildings, I think, would be susceptible to strong ground shaking."
Kelman, approaching the same point from the disaster governance side, notes that Iran's earthquake engineers and seismologists are "among the best in the world, but have not had governmental support." The technical knowledge exists. The political will to act on it has not.
What the War Changes — and What It Doesn't
Both researchers are careful about the compound risk claim. "I think the chances of a major earthquake in Iran in any four-week period are pretty low," Nissen says. "The chances over a ten-year period are pretty high." The argument is not that an earthquake is imminent. The argument is about what an earthquake would cost if it happened — and whether the models pricing that risk have kept pace with what the war has done to the city around the fault lines.
Kelman offers the more cautious take on the war's direct impact. "The amount of infrastructure damage which has occurred because of the conflict is a small fraction of what would be expected due to the earthquake." The baseline losses, he argues, are already so catastrophic that the marginal effect of wartime degradation is relatively small. But he is unambiguous about one dimension the war has changed: the international response.
"We would expect to see that the immediate bombardment would stop and that the world would certainly be very responsive," Kelman says of a hypothetical earthquake scenario. But he adds that what happens after the immediate emergency phase is historically predictable and not encouraging. Iran has had three major earthquakes — in 1990, 2002, and 2003 — that all generated hope of diplomatic opening. None produced lasting change. After Bam in 2003, Iran expressly refused Israeli aid "even though the most experienced and best urban search-and-rescue teams in the region were coming from Israel," Kelman notes. "The Iranian regime made a specific decision that they would rather that lives be lost than accept aid from Israel."
Nissen's worry is more practical and immediate. "What I would really worry about is the first response to an earthquake — are first responders able to get in there in the way that they normally would? Are the medical facilities functioning? Are the roads allowing first responders in, or is that being damaged by bombing?" He adds: "Time is of the essence. People buried in rubble can survive up to a few days if they have water. If they don't, they will die within something like three days." A degraded road network, a damaged hospital system, a civil defense apparatus stretched by conflict — each compresses that window. "I think there are definitely compounding risks," he says.
The Gap in the Models
For the risk markets, both Kelman and Nissen surface the same structural problem: the models used to price Iranian earthquake exposure were calibrated to peacetime conditions and have not been updated to reflect the changed environment. "The models are very poor" when it comes to how conflict intersects with disaster, Kelman says, partly because "so much regarding how conflict impacts disasters is up to personal choices — the choices of leaders, the choices of people with power."
Commercial catastrophe models treat emergency response capacity as a static baseline input. They ask what the earthquake does to the buildings. They do not ask what the earthquake does to a city whose hospitals are partially offline, whose fuel distribution is disrupted, whose civil defense coordination has been stretched by weeks of conflict. The February 2023 earthquakes on the Turkey-Syria border offered a partial natural experiment: the same geological event produced substantially higher mortality per unit of ground shaking on the Syrian side of the border, where a decade of conflict had degraded the response infrastructure, than on the Turkish side.
The gap between what the models show and what an earthquake would actually cost is not theoretical. It is active, and it widens with each passing week.
"Someday there's going to be a big earthquake in Tehran," Nissen says, "and it's going to be cataclysmic."
What Oil Price Models Say About the Hormuz Crisis
The closure of the Strait of Hormuz represents something genuinely new in the history of oil market disruptions, not just in its geopolitical character, but in its scale.
Roughly 20% of world oil supply transits the strait and that's approximately twice the supply loss associated with either of the two closest historical analogues: the 1956 Suez Crisis, which disrupted about 10% of world oil output, and Iraq's invasion of Kuwait in 1990, when the two countries together accounted for about 9% of world production.
That framing comes from Christiane Baumeister, a professor of empirical macroeconomics at the University of Notre Dame and one of the leading researchers in oil price forecasting. In a recent conversation for Risky Science Podcast, she walked through how her structural modeling framework applies to the current crisis — what history can tell us, where the genuine unknowns are, and what her models say about the downstream effects on inflation and growth.
Two Precedents, One Structural Insight
Baumeister draws on two historical episodes as the closest comparators.
The Suez Crisis of 1956–57 is instructive because of the geography of dependence: Europe at the time sourced roughly 70% of its oil from the Middle East and absorbed the brunt of the disruption, a position analogous to Asia's today, given that most Hormuz shipments are bound for Asian markets.
The 1990 Iraq-Kuwait conflict is relevant for a different reason: it involved the same cast of countries now implicated in the current crisis, including fears of contagion spreading to Saudi oil infrastructure, a risk that's once again in the market's line of sight.
In 1990, crude oil roughly doubled within a couple of months. The spike was ultimately contained because Saudi Arabia had substantial spare production capacity and deployed it. That's one of the key structural differences to assess today: the US, now the world's largest oil producer, is a potential counterweight — but how shale producers will actually respond is, in Baumeister's assessment, the genuine wild card.
"They got burned a lot during COVID," she noted. "They recovered, but the question is what their objectives are right now and whether they feel compelled to step in and make up for the shortfall."
The GECO Indicator and What It's Watching
Baumeister regularly updates her Global Economic Conditions (GECO) indicator, which aggregates transportation activity, real economic output, financial market signals, and uncertainty measures into a single composite. Its value in a crisis isn't just the headline reading — it's the decomposition. By isolating which components are moving, the indicator can identify where disruption is actually transmitting through the economy.
During COVID, transportation dominated. Currently, as of her February update, financial market developments and real economic activity were the key drivers. She expects the next update to show significant movement across the geopolitical risk index, oil price uncertainty measures, consumer confidence, commodity prices, and financial variables — all of which have been volatile since the closure of the strait.
She also flagged something notable: in her assessment, markets have so far been somewhat underreacting to the longer-term uncertainties the Hormuz closure implies.
Supply, Demand, and the Amplification Mechanism
A central theme in Baumeister's research is that the source of an oil price shock matters as much as its size. Supply shocks and demand shocks have different downstream macroeconomic profiles, and conflating them produces misleading forecasts. Her structural framework uses economic knowledge about producer and consumer behavior — along with storage incentives and activity indicators — to separate the two in real time.
The current episode is clearly supply-driven. But she identifies uncertainty not as an independent additional shock, but as an amplification mechanism layered on top of the supply disruption: it deepens the real economic consequences without being a separate causal force.
Precautionary demand is also a factor. When buyers anticipate prolonged shortages, they accumulate inventory now, adding apparent demand to a market already under supply pressure.
👉 Listen to the full episode
