A Record El Niño Is Coming: How One Weather Pattern Bends the World's Growth Trajectory

Pitiful — the clothes on his back are thin, yet he dreads cheap charcoal, and prays for colder weather.

— Bai Juyi, “The Old Charcoal Seller,” Tang dynasty

Twelve hundred years ago Bai Juyi wrote about an old man selling charcoal: dressed too thinly for the cold, yet praying the cold would deepen, because warm weather meant his charcoal wouldn’t fetch a price. Weather wasn’t the backdrop to his life. Weather was his income statement.

This episode is the same story, scaled up to the whole planet.

What the episode is about

Odd Lots hosts Justin Mankin, professor of geography at Dartmouth and director of a research group that estimates the economic damage of climate. The subject is the El Niño now forming — one that every forecasting model currently points toward being potentially record-breaking.

What makes the opening fun is that the hosts start out bickering. Tracy calls herself a weather nerd; Joe calls himself a wind chill truther — it’s 25 degrees out, you tell me it “feels like 15,” how do you know what 15 feels like? The guest promptly dismantles that in his first answer: wind wicks moisture off your skin, and the phase change from liquid to vapour pulls latent heat away. That’s a measurable physical quantity, not a vibe. Joe later admits he went 45 years without knowing what El Niño actually was, and now he gets it.

That exchange is worth keeping, because it sets the tone. A term everyone has heard and assumes they understand turns out to be something else entirely. And the real payload comes later — not how bad this one will be, but that we may have been computing “bad” in the wrong units all along.

The main points

1. The physics: warm water sloshes east, and the whole room heats differently

In normal conditions, easterly trade winds drag sun-warmed surface water westward along the equator, piling it into a deep pool of warm water parked off Indonesia and Australia. Every few years those trade winds weaken and break down, and that pile spreads itself out across the vast expanse of the tropical Pacific.

Mankin’s analogy is the best line in the episode. Neutral or La Niña conditions are like walking into a cold room with a small space heater plugged in one corner. El Niño is the same room with wall-to-wall radiators instead. The energy hasn’t increased — but redistributed, the thermodynamics of that room are a completely different experience. The tropics are the heat engine of global weather, and spreading that warm water out exposes far more of the atmosphere to that heat.

2. Why this one is different: two unrelated methods converging on the same answer

There are two families of El Niño forecast. Statistical models take temperatures at time t and project forward using autocorrelation in the series. Process-based models actually simulate the coupling between ocean temperature, winds, pressure and circulation. The two share almost no assumptions — and both are currently converging on a record event, at least since 1950, and warmer than the 2015–16 El Niño. Australia’s relative oceanic Niño index, per the hosts, is projected to reach a level it has never previously touched.

One date worth holding onto: these events typically peak in December through February, and this one isn’t expected to wind down until February or March of 2027. Mankin says it has already nearly locked in a record annual temperature for 2027 — a year that hasn’t happened yet, but whose books are already partly written. The hottest years on record have almost always been El Niño years, because you’re imposing an anomaly on top of a warming trend.

3. The core finding: not a collision, a trajectory bent downward

This is the heart of the episode. The 1997–98 El Niño was estimated at the time at roughly $36 billion in economic losses — a number computed as a level effect: a shock, then a recovery.

Mankin and his collaborator Christopher Callahan redid that accounting using causal inference techniques and got something entirely different. El Niño appears to systematically depress growth. Your country was growing along one trajectory; the event happens; afterwards you are growing along a different, lower one. Accumulate that over five years and the single 1997–98 event amounts to roughly $5.7 trillion in global losses by 2003.

The gap between $36 billion and $5.7 trillion isn’t an arithmetic error. The two numbers measure different things: one asks how much fell over during the impact, the other asks how much growth never happened after the trajectory was bent.

4. Why the trajectory bends: global simultaneity

Mankin’s explanation is simultaneity. A heat wave or drought in the upper Midwest tells you nothing about Russian grain, or about cotton prices in Uzbekistan’s Fergana Valley. Those shocks are disparate enough in space that they can offset each other. El Niño isn’t like that — it generates many different kinds of hazard at once: heat waves, droughts, floods, landslides, wildfires. Your supply chain isn’t disrupted in one place, it’s disrupted in a dozen. When many regions are absorbing losses at the same time, the efficiency with which capital and labour get mobilised into productive parts of the economy degrades.

He gives one very concrete cascade. The onset of the South Asian monsoon determines feast or famine for billions of people, and this El Niño has already delayed it. A delayed monsoon means farmers who wouldn’t otherwise have irrigated now must — and that irrigation runs on diesel generators, so diesel demand and prices spike. Diesel subsidies are a live wire in Indian politics, and local politicians’ electoral fortunes ride on them. A sea surface temperature anomaly ends up landing as the outcome of a municipal election.

5. He marks the boundary of his own estimate

For the event now unfolding, Mankin’s figure is roughly $10 trillion in global economic losses over the next five years, conservatively — up to $14 trillion if you include some of the tail scenarios. But he immediately adds a caveat: only about 70% of that is in sample, meaning supported by existing data. The rest is extrapolation.

There’s a second mechanism that’s easy to miss. The size of the losses depends not only on the magnitude of the El Niño but on the size of the global economy. In his words: there’s just more stuff to be impacted. An identical event today produces a larger headline number than the same event in 1997, for reasons that have nothing to do with the weather.

6. Peruvian farmers were forecasting El Niño five hundred years ago

The most human moment in the episode. Work by anthropologist Ben Orlove with climate scientists Mark Cane and John Chiang found solid evidence that Andean potato farmers were forecasting El Niño as long as 500 years ago — by the visibility of the Pleiades.

The physics genuinely holds up. El Niño shifts convection that normally happens near Australasia eastward, generating a shield of high cirrus cloud over the Andes, which dims that star cluster. The farmers saw the stars go hazy and adjusted their sowing dates.

Mankin argues the highly industrialised world, still leaning on instrumentation, has underappreciated how much cultural knowledge of this kind exists around the world. It’s worth pausing on: that isn’t folklore. It’s a leading indicator with a real physical chain behind it. The instrument just happened to be a pair of eyes.

7. The sharpest line: centuries of exposure, and we still haven’t adapted

The hosts ask the obvious question — countries get richer, infrastructure hardens, forecasts improve, firms learn to anticipate. How do you model all that adaptation?

Mankin’s answer is blunt: the evidence is pretty mixed, and there isn’t a whole lot of evidence that we’re well adapted to the climate we already have. His reasoning is compact. We’ve been exposed to El Niño for hundreds of years — this is not a novel phenomenon — and yet the measured costs remain astounding. That fact alone tells you we are not optimally adapted.

He does describe what adaptation looks like. In 1982–83 there was almost no lead time to plan for the impacts, and the value of a forecast is precisely the lead time it buys. Beyond that: early warning systems, emergency management teams positioned where they’ll be needed, and insurance markets that actually recognise the region-specific geophysical hazards involved. He also separates two questions that usually get muddled — how you make international markets resilient to shocks, versus how you keep an individual safe from a hazard. They need different tools, and he says simply pulling them apart is itself valuable.

Going further

”The news says a record weather event is coming — should I adjust my portfolio?”

Probably the first thought many listeners have. Start with a clue the episode hands you: Tracy mentions that in just the past few days she’d seen a Citi note on Thai interest rates, a Morgan Stanley outlook on the Bank of England, and a Bank of America FX outlook — all referencing El Niño.

When sell-side notes start collectively naming the same meteorological phenomenon, what you’re looking at usually isn’t new information. It’s a consensus that has already been priced once. That doesn’t make it false; it makes knowing it, by itself, not an edge.

The deeper issue is timescale. The episode’s central finding is about a growth trajectory, not a price level. If the effect really is a depressed growth rate rather than a bounce-back shock, then it manifests as: for several years, every quarter comes in slightly below where it otherwise would have. That does not produce an identifiable candle on any chart. Trying to trade this week off it means taking a signal denominated in five-year growth accounting and using it to answer a question denominated in days. The units don’t match.

So what question does match? Treat it as a change in the denominator rather than an event in the numerator. The question isn’t “what should I buy,” it’s “do the growth assumptions embedded in what I already own quietly presume a world that keeps running as normal?” The first requires you to forecast weather. The second only requires you to audit your own assumptions — and that one is actually doable.

And state the falsification condition plainly: if this event lands inside the historical range and supply chains don’t break in a dozen places at once, then none of the above happened. You don’t get to go back afterwards and say it mattered anyway. That would turn a refutable inference into a claim that’s always right and therefore useless.

”Ten trillion, fourteen trillion — the numbers are too big to feel. How do I know what to believe?”

The episode demonstrates a reading method you can lift directly, and the guest hands it to you himself.

First layer: he volunteers that only about 70% of the estimate is in sample. An estimate willing to say which part of itself has no data behind it is signalling quality, not weakness. Inverted: a number presented with no boundary, no account of where the uncertainty comes from, is the one to be wary of. That habit transfers cleanly to research reports, price targets, and every “the market will reach X by 2030” slide you’ll ever see.

Second layer: the phrase “record-breaking.” The episode notes the index itself is benchmarked against a rolling baseline, because the tropical Pacific is warming secularly underneath. Which means part of “record” comes from how the baseline was defined. That isn’t an accusation of dishonesty — it’s a reminder that every “highest ever” or “lowest ever” claim requires you to ask “compared to what?” first. The same discipline applies to earnings and to backtests: if a new high dissolves when you shift the base period, it’s a slogan, not a finding.

Third layer, and the part I’d most want to keep. Mankin describes a “perfect model” exercise. They didn’t ask “is the effect permanent?” They asked the inverse: suppose it truly were permanent, and of exactly this magnitude — could we recover that known effect from the data record we actually have? The answer was no. So their conclusion isn’t “proven permanent.” It’s “we cannot rule out that it’s permanent.”

What that move does is separate “I don’t know” from “it isn’t there.” Most research runs the other way: pick a conclusion, then gather evidence for it. This test measures the length of your own ruler before deciding how strongly you’re entitled to speak. If your data can’t reproduce an answer you already know, it also isn’t qualified to rule anything out on your behalf.

”I’m already diversified — surely global risk is covered?”

The episode challenges that intuition directly, and the reasoning is solid.

Diversification works on the premise that the shocks hitting your positions are uncorrelated. El Niño’s damage comes precisely from simultaneity. When one factor hits agriculture, mining, transport, insurance and energy at the same moment, the diversification you thought you had collapses into a single position for the duration. It’s the same feeling as 2020: things that normally offset each other all move the same way inside a genuinely global shock.

Mankin’s own framing when asked about defence is telling. He admits he doesn’t know much about how people build portfolios, though he knows diversification is a strategy for securing returns. But the distinction he draws next is more useful than any allocation advice — he separates “how do you shore up international markets so they’re resilient” from “how do you protect a person from a hazard,” and says that simply pulling those two apart is really valuable.

That transfers exactly to the personal level. Portfolio resilience (correlation, position size, liquidity) and life resilience (cash buffer, insurance, stability of employment income) are two different systems. Using one to answer questions belonging to the other is how you end up looking beautifully diversified on paper and unable to survive a month without income.

Which brings it back to the sharpest line in the episode: centuries of exposure, and the costs remain astounding. Time by itself does not make you more able to absorb risk. Only actually taking the feedback on board and changing what you do does that. The test on yourself is simple enough: if your reaction, your position sizing, and what went through your head were identical in the last three drawdowns, you haven’t accumulated experience. You’ve been exposed three times.

Sources worth checking

• Odd Lots, Bloomberg, episode of 14 August 2026, with Justin Mankin, professor of geography at Dartmouth College
• Christopher Callahan and Justin Mankin, “Persistent effect of El Niño on global economic growth,” Science, 2023
• Ben Orlove, John Chiang and Mark Cane, on Andean farmers forecasting rainfall and crop yields from Pleiades visibility, Nature, 2000
• Mark Cane and Stephen Zebiak, the first physical forecasting model of El Niño, mid-1980s
• Mike Davis, Late Victorian Holocausts, Verso, 2001, on the 1877–78 El Niño and the famines in India, northern China and northeastern Brazil
• NOAA Climate Prediction Center, ENSO advisories and outlooks
• Australian Bureau of Meteorology, ENSO Outlook and the relative oceanic Niño index

Disclaimer

This is a personal reflection on a publicly available podcast episode, offered as educational commentary and note-taking. It is not investment advice and does not constitute an offer, solicitation or recommendation regarding any financial instrument. No price targets are given and no individual security is recommended. All figures and claims are paraphrases of the episode or citations of public sources and may contain transcription or interpretation error; please consult the original sources. Markets carry risk. Any investment decision should rest on your own financial situation, risk tolerance and independent judgment, or on the advice of a qualified professional.

Raw Markdown

This article is an educational discussion of investment method. It is not advice to buy or sell any individual security, offers no target prices, and does not analyze any current holding. Investing carries risk; make your own decisions or consult a qualified professional.