Adam Rozencwajg — The Oil Supply Shock Hiding in Plain Sight
"Just because we haven't had a crisis yet doesn't mean it won't come." Six months of a 10 mb/d shut-in, 400M+ barrels already drawn, and the shortage quietly migrating from crude into refined products.
One-line take: The market has decided G&R got the Hormuz call wrong because no crisis has shown up — and Rozencwajg argues that is exactly the wrong inference. 10 mb/d of upstream production has been shut in since early March; on his May math 100 days of that is a billion barrels that never reached market, and the world never had a billion barrels of drawable inventory — most reported stock is working capital filling pipelines and tankers, not a savings account. Three things have masked it: a 45–50 day lag before the missing barrels showed up (only three of five months of data have felt it, so even a reopening tomorrow means two more months of draws); a one-time ~100M-barrel flotilla that escaped the strait when the MOU was signed; and — the important one — China, Russia and the Gulf all cut refinery runs, so the shortage transferred from crude into refined products, the half of the balance nobody instruments. ~400M+ barrels already drawn; the host's arithmetic puts the running deficit at 1.76B barrels (day 176 × 10 mb/d) and a non-reckless SPR tank bottom ~18 weeks out. The tells: record 3-2-1 and diesel cracks (~$102 vs a normal $20–30) — diesel priced as if oil were $155 — while the reported "demand is down 5 mb/d" is an artifact of refinery runs as a demand input, contradicted by commercial flight traffic up 5–6% y/y. Behind the cycle: shale peak growth was 2019 (his own neural-net call), the Permian is down to six counties and gassing out (the "gas burp"), and the industry spends ~$550B/yr against the ~$1.5T/yr needed to recapitalize — while boardrooms still prefer buying back stock at half NAV to drilling. Positioning: long the crude molecule through the equities — deliberately not the US refiners — plus a small, not-yet-full-conviction fertilizer add (sulfur $300→$1,100/t; Qatari LNG and phosphate both moving through Hormuz).
1. Stocks & names mentioned
| Ticker | Name | Research | View | What he said | At |
|---|
| Crude oil | Crude oil (the molecule) | — | Positive | "I continue to think that the oil molecule is the mispriced asset." The primary mover of the crisis is 10 mb/d of upstream production turned off; the catch-up trade "is oil up to diesel," and refilling drained gasoline/diesel/jet tanks means "we're going to have to bid a lot of crude back into the refining system, and that's going to catch people off guard." | 51:03 |
| Oil producers | Oil & gas producer equities (E&P) | — | Positive | How G&R expresses the call: "I'm not a physical or paper commodity trader. I trade the equities." The stocks "didn't do very much throughout this crisis" and "remain out of favor" because sentiment never turned — while the sector needs ~$1.5T/yr of spending vs the ~$550B it does, and management teams still prefer buying back stock at half NAV to drilling. | 39:58 |
| Nat gas / LNG | Natural gas & LNG | — | Positive | "All the LNG coming from Qatar has been disrupted, so the LNG gas market is very very tight." US supply has held up only because of the Permian "gas burp" — an aging-field artifact — "and that's the sign of a field that's getting older… this kind of last gasp of gas." Everywhere else US gas "has been quite weak." | 32:10 |
| Fertilizers | Fertilizers (phosphate, ammonia / urea) | — | Positive | "We increased our fertilizer exposure a little bit, but it's not yet to the point where we're ready to make it a full conviction." Availability is "extremely tight": phosphate transits Hormuz, ammonia and urea are made from natural gas, and Qatari LNG is disrupted — "so it's going to be tough to get those two fertilizers on." | 1:00:40 |
| MOS | Mosaic | QT · SA · STK · FA | Neutral | Raised by Martenson, not rated by Rozencwajg: Mosaic "cut back their phosphate production pretty hard," probably on the sulfur price, which "went from 300 to 1100 a ton" — the one price signal in this crisis that "makes sense." Rozencwajg's answer is the sector view, not the stock: fertilizer exposure up "a little bit," not full conviction. | 1:01:01 |
| Diesel | Diesel & refined products | — | Neutral | Physically the tightest thing in the complex — "refined product inventories are probably falling faster than anyone realizes," maybe 4 mb/d uncaptured — but already richly priced: the diesel crack at ~$102 vs a normal $20–30 means "diesel's trading as if oil was 155 a barrel." If he traded paper, "I probably would be betting that the crack spread would come down with oil benefiting." | 49:26 |
| Grains / ag | Grains & agriculture | — | Neutral | "The jury is still out." A monster El Niño historically brings volatile weather but also wetter summers, which "could actually be somewhat bearish for crop prices" against a global drought — yet the crop is "priced to perfection" after 15–20 years of unbelievable yields with demand rising: "it's a little bit like a trade where everyone's on one side of the boat." A tail risk, not a position. | 1:00:14 |
| Saudi Aramco | Saudi Aramco (Abqaiq processing facility) | — | Neutral | Infrastructure data point, not a stock view: "Abqaiq, the oil processing facility, was hit again a couple weeks ago, which went fairly under-reported — the same attack that happened in 2019." He is nonetheless "a little less worried" that shut-in Gulf fields won't restart, since they came back cleanly after COVID and after the MOU reopening. | 55:22 |
| US refiners | US refiners (as a class) | — | Negative | Deliberately avoided: "we didn't invest in the US refiners, for instance… we thought that the crude oil molecule was the mispriced asset." Record cracks are the symptom, not the prize — the shortage began upstream, and his implied bet is that "the crack spread would come down with oil benefiting." | 53:45 |
"View" is Adam Rozencwajg's stance in this conversation (Positive / Neutral / Negative), not a price rating. Research links: QT Qualtrim · SA Seeking Alpha · STK Stock Analysis. This is a physical-market conversation — crude, products, gas, fertilizer and grains are discussed as commodities and asset classes, so most rows carry no ticker. Mosaic is the only listed company named (by the host).
2. Talking points
1:49 The setup — Hormuz closed, 10 mb/d shut in at the wellhead
- "Really quiet summer over here… obviously, I'm being facetious." Operation Epic Fury started in early March, the Strait of Hormuz closed, "and that led to the huge oil disruption that we've really been living through ever since."
- 20 mb/d crossed the strait pre-crisis. Some rerouted by pipeline to the Red Sea and overland into Turkey, but "by all measures it's impacted upstream production by about 10 million barrels per day. So 10 million barrels at the wellhead got shut in."
3:15 The billion-barrel deficit — and the two places it can come from
- The May report's arithmetic: 10 mb/d away for 100 days is "essentially a billion barrels of oil that never made it to market." It has to be paid for out of inventories or out of demand destruction.
- 3:54 The catch: "we didn't really have a billion barrels of easily mobilized inventories." Headline stocks overstate it because much of it "is really more working capital" — oil that "fills pipelines and it fills tankers… it's not readily available to be drawn down. Think of it like working capital in your business as opposed to straight savings in your savings account."
- 5:05 Hitting tank bottoms would be COVID in reverse: then we hit tank tops and "oil traded for minus $50 a barrel." Now "you might get a spike up to $2 or $300, might not be sustainable" — the market has gone "from having a buffer to operating a giant global just-in-time energy market, which we've never really done before." JP Morgan's separately-derived math reached the same shortfall.
6:10 The MOU round-trip — "they said G&R got it wrong"
- Third week of August: the tanks haven't run dry, and when the US and Iran signed the MOU "oil prices traded all the way back down to where they were before the war started."
- The perverse inference: the IEA claimed a glut in January, and every day 10 mb/d was disrupted without a crisis became "ammunition for the bears… this proves that the glut must have been even bigger."
7:26 2022 vs 2026 — the positioning was the opposite way round
- 2022 was a bullish mistake: investors bid oil to $125 on 5 mb/d of Russian exports that might be lost, the loss never came, and the year ended flat.
- This year started "hugely bearish" — energy 2.5% of the S&P, gross speculative shorts at extreme levels going into the war weekend.
- 9:18 The Monday spike to ~$120 was risk managers forcing shorts to cover; the gross short on NYMEX and Brent then "started creeping back up and back up" until, by the MOU signing, it was back to pre-war levels. "10 million barrels was disrupted for much longer than anyone anticipated… but the sentiment throughout really hasn't changed. And that's why the oil stocks didn't do very much."
10:40 Hostilities resumed, the strait is shut again — and the market is complacent
- The MOU "has sort of gone away again… the strait is now closed." Yet the universal view is that because there has been no crisis, there won't be one — "a very dangerous and complacent view to have in the market today."
- 11:35 On the "mystery short seller" stories: he found nothing. The paper market is "40 times larger than the physical," but unlike equities the oil market "is grounded on some level in reality" — 105–106 million physical barrels a day, "$10 billion a day at $100 oil" — so "I always prefer to look for an answer that's based in physical volumes."
13:09 Why no crisis yet, #1 — the 45-to-50 day lag
- Inventories didn't start falling until 45–50 days after March 1: first the onshore tanks emptied, then the vessels, then the fields shut in — and boats normally take 20–30 days to reach their destination.
- Of the five months of data (March–July, July preliminary), "two of those five months didn't feel the impact." Symmetrically, "if the strait were to reopen tomorrow… it would take two months to work it back into the system on the other side. So we're only just more than halfway through the physical manifestation of this problem."
- 14:41 "Where we are right now, we've drawn like 400 million plus barrels" — against a billion-barrel nut that "should be even higher than that" because the disruption has outlasted the assumption.
15:38 Why no crisis yet, #2 — the one-time flotilla "burp"
- When the MOU was signed the strait briefly reopened and the backlog of vessels trapped in the Persian Gulf "made a beeline out of there" — roughly "100 million barrels out in fairly short order."
- 16:13 It landed into a market with three refining systems down: Gulf refineries (whose crude was exported rather than run, "let's just get out while the getting's good"), Russia (refining infrastructure destroyed by Ukrainian strikes), and — "the really notable one" — China, which stopped importing, refining and exporting products to concentrate on its domestic market. Why is debatable: policy, or "a dry run for what China might look like if it were all of a sudden embargoed."
18:36 Why no crisis yet, #3 — the shortage moved into products, the market's blind spot
- If end users kept consuming while refiners didn't run, "that just means you've shifted the problem from crude to refined products… refined product inventories are probably falling faster than anyone realizes." And "nobody has a good handle, particularly in the emerging market world, what level are the refined product inventories at today?"
- 19:32 The mechanism that ends the calm: "we have to refill those gasoline, diesel, jet fuel tanks that are probably at dangerous levels… and to do that, we're going to have to bid a lot of crude back into the refining system."
- 19:50 Why the standard models break: supply is crude and NGLs at the wellhead, demand is refined product demand, and refining is normally netted out "like an intercompany transfer." Dislocate that middle and "all of your models that just kind of treat this as one system break apart."
20:38 The flight-tracker demand check
- Martenson's real-time demand proxy is the flight tracker; flights are running above last year. Rozencwajg: "that's a metric that we like to look at as well." It isn't one-for-one (5–6% traffic growth is not 5 mb/d), but "the two series do line up incredibly well," and the correlation held even through COVID and the '08 recession despite jet fuel being ~9% of demand.
- 22:36 "I can say with almost 100% certainty that we are not down 5 million barrels a day, which is what you'll read in the headlines… those two things are irreconcilable to me."
- 23:13 On the agencies: EIA data is better and "largely less editorialized" than the IEA, which releases a bearish chairman's summary to the press hours before the backing data — "always a bearish commentary backed up by some bullish data." Both have systemic errors, "but I'm awfully glad I have it."
24:38 1.76 billion barrels missing — against a forecast that snaps back to normal
- Martenson's running total: outages of 10–16 mb/d against a normal ~2, call it 10 mb/d for 176 days — "we're missing 1.76 billion barrels now." The EIA's August Short-Term Energy Outlook concedes 55 days of crude-and-other-liquids supply cover, then has it "trundle right back up to normal range through 2027."
- "That is really where the rubber meets the road." G&R was bullish before the war because inventories weren't building the way a headline surplus implied — "that told us that the headline numbers were likely incorrect." So rebuilding stocks off dangerously low levels "is going to be very difficult… that's where I think there could be quite a bit of panic," and where the story shifts "from a short-term trade the headline to a longer term, oh, we have a problem."
26:30 The shale glut ended the capital cycle — and nobody restarted it
- Shale gave "basically a glut for 10 or 12 years," and the belief that "if oil got too high, the shales could just bring on more" removed any sense of scarcity — "it results in capital leaving because everyone grows complacent. No one invested in the next generation of fields."
- 27:49 The IEA's base-decline study showed a gap between the ~$550B/yr actually spent and what holding production flat requires — and Rozencwajg argues flat is the wrong target since demand grows, while the IEA's own numbers still assume shale contribution. "On my math you probably need… about a trillion and a half a year in spending over a decade or so to really rehabilitate and recapitalize the energy industry."
29:45 The Permian "gas burp" — gas-oil ratio as a field-aging tell
- US crude production has "essentially turned negative year on year," so the EIA's projected +400 kb/d for 2027 "can only come from natural gas liquids."
- The geology, in his soda-can analogy: gas is dissolved in the oil in the same wellbore; as reservoir pressure falls "the gas separates out from the liquids… you get a whoosh of gas that comes out. And that's a sign that that field is getting older." Drilling growth kept the average well young; now the field matures, "the average age of the well is getting older and older… so it's becoming gassier."
- "We got this wrong… we didn't have that modeled properly a couple years ago." It explains why US dry and wet gas "has been able to hang in there better than we would have expected" — but it is "this kind of last gasp of gas."
- 32:36 And NGL growth is rolling over too: "if gravity and inertia is pulling all these streams down, yes, the split to gas goes up and so you can get this little dislocation, but it doesn't last."
33:19 "We have plenty of oil" — versus a country that is a net importer
- Reacting to a clip claiming oil never hit $300 because "we have a lot coming up" from Texas, Alaska and Louisiana, Martenson notes the EIA's weekly report shows the US as a net importer of crude: "how does a net importer of oil supply the world?"
- 34:40 Rozencwajg grants the long-run point — the shales did break the 2007–08 Malthusian scare, and "as they say in our business, the cure for high price is high price" — the US is "unbelievably well-endowed." "But the question is just the path in the middle," and a new price and capex cycle "takes years and years."
- 36:33 "The only one left is the Permian. You're down to six counties in West Texas" — he sees it directly on the royalty/leasing side as brokers bring him increasingly fringe acreage. The 2019 call, made with deep neural nets G&R built before public LLMs existed: shale was growing 2 mb/d with growth still accelerating, and they called peak growth, not peak production — never to exceed 2019, turning negative year-on-year by 2025. "That's exactly what's happened."
38:39 The boardrooms still won't drill — capex trims, an ANWR flop, and buybacks at half NAV
- Despite the administration's "drill baby drill" arrows, growth hasn't come: prices are high, regulation isn't the binding constraint, "productivity has just trended lower and lower" — so the talk has shifted to asking Saudi Arabia to produce more. Meanwhile E&Ps are trimming capex budgets and the ANWR auction "was a flop, a complete flop."
- Why: budgets are set off the futures curve, which "hasn't done nearly as much as the spot curve has," so incentive prices look poor, and "oil stocks remain out of favor."
- 40:19 The arithmetic G&R ran ~5 years ago: cash on the balance sheet is credited at 100 cents on the dollar; spend $100M to create $200M of NPV and the market capitalizes it at half NAV — so the value is unchanged, you've lost the cash and drilled your best acreage. "But of course, there was another option… you buy back your stock at half NAV. Now all of a sudden that was massively accretive" — the disincentive to drill that "has lingered in the market."
42:10 SPR tank-bottom arithmetic — roughly 18 weeks of drawdown left
- Martenson's build from a 170-page DOE report: 331M barrels at the time of the analysis (now ~290M), less 130M in single-cycle early storage caverns "nobody's that reckless" to drain, less a 10% heel (~58M) on the remaining capacity — leaving ~143M then, ~110M now, and at 6M barrels a week "about 18 weeks from now" to a non-reckless tank bottom.
- Rozencwajg: "I think that's fairly good math." The OECD's announced 400M-barrel release was a Draghi "whatever it takes" / Bernanke-2008 gesture, and "the government didn't release at the initial pace that they claimed to," which suggests 400M is about the comfort limit.
- 45:07 "It would be the height of stupidity, the absolute height of stupidity, to not look to refill strategic petroleum reserves after what we've learned here… it's a very very cheap insurance policy."
45:36 The product picture — and the missing 5 mb/d of "demand"
- Commercial crude stocks look fine ("right in range"), but gasoline and distillate are 10–15M barrels below where they should be, Cushing "is just dragging the bottom," and Middle East plus Russian gas-oil/diesel exports have "absolutely cratered… down 80% from its usual 2.5 million barrels per day."
- 46:42 The balance: 10 mb/d of supply lost, ~4–5 mb/d covered by inventory draws, and the residual booked as a 5 mb/d demand collapse — bigger than 2008 and comparable to lockdown COVID. "And we just don't see it… the airports are absolutely packed." The reason the number exists: "one of the big inputs in your model to estimate demand is refinery runs. So if the refiners stop running not because of an economic slowdown, but because of a war, you could see that as a signal that demand is down. And that's the tail wagging the dog."
48:11 Crack spreads are the diagnostic — diesel priced as if oil were $155
- If demand really were being rationed away, "oil should be relatively weak and refined product prices should be relatively weak" together. Instead "the 3-2-1 is at record highs and just straight gross diesel pricing is at record highs" — which says demand is not impaired, and roughly 4 mb/d of refined-product inventory is draining uncaptured "because we have such bad data on refined product inventories."
- 49:49 The diesel crack at ~$102 against a normal $20–30 refining margin puts ~$70 of excess on ~$85 crude — "diesel's trading as if oil was 155 a barrel." Rozencwajg agrees that is where it should be, and adds the demand-destruction benchmark: the only two real episodes were 1981 and 2008, "and in both of those cases in real dollars adjusted for today, oil was like close to 200 bucks."
- 51:03 "The big catch up that's going to take place is oil up to diesel" — the only things holding crude down are the one-off flotilla and China "essentially transmitting the crude problem… keeping the price of crude low at the expense of the price of refined product." Somebody will optimize other refineries to collect a spread that size.
53:45 Own the molecule, not the refiners
- The hypothetical of knocking out all refining (products to infinity, crude to zero) "misses the point in the real world. The issue at the end of the day is that you've taken 10 million barrels a day of upstream production offline." Russia and China are layers on top; the primary mover is "the fact that you turned off the taps upstream."
- "So we didn't invest in the US refiners, for instance… we thought that the crude oil molecule was the mispriced asset today. I continue to think that the oil molecule is the mispriced asset." He trades the equities, but if he traded paper he'd bet "the crack spread would come down with oil benefiting."
- 54:41 On whether shut-in Gulf fields restart: less worried than the consensus — whole countries were shut in during COVID and "they came back online no problem," and a few producers even raised field-level output during the brief MOU reopening. Abqaiq was hit again "a couple weeks ago… which went fairly under-reported." Any restart failure is "a bullish surprise as opposed to baking it into my estimates."
56:37 The polycrisis behind the oil — sulfur, urea, LNG, El Niño and a global drought
- Martenson frames 2027 as "spicy," especially around food: urea and sulfur missing out of the Gulf, plus a monster El Niño.
- El Niño cuts both ways — historically more volatile weather (bad for yields) but also wetter summers, which against today's "big widespread drought," very low groundwater and dry soil "could actually be somewhat bearish for crop prices."
- The clearer channel is fertilizer: "phosphate goes through the Strait of Hormuz, and then ammonia and urea is made through natural gas conversion, and of course all the LNG coming from Qatar has been disrupted." 2026's crop looks okay, "but we didn't get enough fertilizer on the fields this year either, already."
- 59:15 The drought is global — the US, "all of Europe and the breadbasket and Ukraine… the Black Sea," and Brazil, so a wetter US could be offset by a drier Brazil; a failed monsoon puts India "obviously big trouble."
1:00:14 Agriculture is priced to perfection — and the fertilizer position is deliberately small
- Fifteen to twenty years of rising protein-driven grain demand has been met by "these unbelievable crop yields… it's a little bit like a trade where everyone's on one side of the boat." Ask instead "how vulnerable are you to even a minor disruption in growing conditions" — including lack of fertilizer. "Volatility in general I don't think is your friend in this type of a market."
- "We increased our fertilizer exposure a little bit, but it's not yet to the point where we're ready to make it a full conviction."
- 1:01:01 Martenson: Mosaic "cut back their phosphate production pretty hard," likely on sulfur, which "went from 300 to 1100 a ton" after a ~50% disruption of globally exported supply — "that price signal, I was like, that makes sense. Not a lot has made sense to me in the energy markets yet."
1:02:08 "Just wait" — the eye of the storm, and the four-more-weeks admission
- The core error he sees: reading the absence of a crisis as proof of abundance. "That's just really really really myopic thinking. You could very well be in the eye of the storm here where it calms down a little bit before things get a lot worse."
- The tell nobody priced: when Trump and Vance were asked why they signed an MOU widely judged a good deal for Iran, "they both said, look, we had to do something because if this crisis lasted even four more weeks, there was going to be a major problem." That was six weeks ago — "and the strait shut again."
- 1:03:05 His falsifiable line: "if we don't have a crisis by, I don't know, November or December… next time we speak, we can try to figure out what's going on." Until then, "even if it were to ameliorate tomorrow, I think the damage might already have been done."
3. In plain English
A jargon-free summary of the thesis behind each call — what it actually is and why he holds that view. (Plain-language companion to the table above; renders on each name's consolidated page.)
Crude oil — the molecule Positive
About 10 million barrels a day of oil — roughly one barrel in ten the world uses — has been stuck in the ground since early March because the tankers that carry it can't get out of the Persian Gulf. Over six months that adds up to well over a billion barrels of oil the world consumed but never produced. That gap has to be filled from storage, and Rozencwajg's central point is that the world does not actually have that much oil it can use: most of what gets reported as "inventory" is oil physically sitting inside pipelines and ships, which you can no more spend than a shop can spend the goods on its own shelves.
So why hasn't anything broken? Three reasons, all temporary. It takes about two months for missing barrels to show up in the data, so only three of the five reported months have felt it at all. A one-off convoy of about 100 million barrels escaped during the brief peace deal. And China, Russia and the Gulf all cut back their refineries, which quietly moved the shortage out of crude oil and into gasoline and diesel — where nobody measures inventories properly.
His conclusion is that crude is the cheap thing in the whole chain. When the world has to refill its emptied fuel tanks, refineries will have to bid hard for crude, and "the big catch up that's going to take place is oil up to diesel." He is not calling a trade on tomorrow's headline — he's saying the physical damage is already done and simply hasn't shown up yet.
Oil E&P equities — oil & gas producers Positive
This is how G&R actually owns the view: "I'm not a physical or paper commodity trader. I trade the equities." An E&P — exploration and production — company is simply a business that drills wells and sells the oil, so its shares are a leveraged way to own the barrel without touching futures.
The appeal is that these shares never went up. Through six months of the worst physical supply disruption in decades, "the oil stocks didn't do very much" and "remain out of favour," because investors decided early that the crisis was a false alarm and never changed their minds. Meanwhile the industry is spending roughly $550 billion a year when, on his math, it needs about $1.5 trillion a year for a decade just to rebuild what's depleting.
The catch is that the companies themselves are part of the problem. Because the market values a producer at about half what its oil in the ground is worth, spending cash to find more oil destroys value on paper, while buying back its own cheap shares creates it — so managements hold production flat and buy back stock. That keeps supply tight, which is bad for consumers and good for anyone who already owns the shares.
Natural gas & LNG Positive
LNG is natural gas chilled into a liquid so it can be shipped. Qatar is one of the world's biggest exporters and its cargoes sail through the Strait of Hormuz — so the same blockage that trapped the oil has made the global gas market "very very tight." That matters beyond heating and power: ammonia and urea fertilizer are literally manufactured out of natural gas, so expensive gas means expensive food inputs.
The subtler half is about American supply. US gas production has held up better than G&R expected, and Rozencwajg says they now understand why — and it isn't healthy. In the Permian basin, gas is dissolved in the oil like fizz in a can of soda. As the wells age and reservoir pressure drops, the gas "whooshes" out ahead of the liquid. So today's steady US gas output is partly a symptom of ageing oil fields, not of new drilling — "this kind of last gasp of gas." He expects it to fade.
Fertilizers — phosphate, ammonia & urea Positive
Fertilizer is the quiet second casualty of the Gulf blockade. Phosphate rock ships through the Strait of Hormuz; ammonia and urea are made from natural gas, and the Qatari LNG that supplies much of that gas is disrupted too. Sulfur — a by-product of refining, needed to process phosphate — went from $300 to $1,100 a tonne after roughly half the world's exported supply vanished. So the three main crop nutrients all got harder and dearer to obtain at once.
Rozencwajg has acted on it, but only lightly: "we increased our fertilizer exposure a little bit, but it's not yet to the point where we're ready to make it a full conviction." The reason for the hesitation is that the demand side — crop prices — is genuinely uncertain: a monster El Niño could bring wetter, better growing weather even as fertilizer gets scarce. Worth owning some; not yet worth betting the fund on.
MOS — Mosaic Neutral
Mosaic is one of the largest producers of phosphate and potash crop nutrients. It comes up as evidence rather than as a stock pick: the host notes Mosaic "cut back their phosphate production pretty hard," and suspects the reason is sulfur, which you need to turn phosphate rock into usable fertilizer and whose price has more than tripled.
Rozencwajg doesn't rate the shares. His answer is at the commodity level — fertilizer availability is genuinely tight, and G&R has added a little exposure without full conviction. The signal worth keeping is the behaviour: a major producer choosing to make less fertilizer because its input costs have exploded is a supply cut in a market the world is already short of.
Diesel & refined products Neutral
A refinery buys crude oil and sells petrol, diesel and jet fuel; the difference between the two is called the "crack spread," and it is normally about $20–30 a barrel. Right now the diesel crack is around $102 — meaning diesel is priced as though crude were $155 rather than the $85 it actually trades at. That is the market screaming that the shortage is in fuel, not in oil.
Rozencwajg thinks that reading is correct: with China, Russia and the Gulf all refining less while people kept driving and flying, perhaps 4 million barrels a day of product inventory is draining without anyone measuring it, and "that's a gigantic problem." Diesel matters more than the others because it moves freight, farm equipment and construction — it feeds into the price of nearly everything.
But being right about the shortage doesn't make it the thing to own, because the price already reflects it. If he traded the physical market, "I probably would be betting that the crack spread would come down with oil benefiting" — that is, the gap closes by crude rising to meet diesel rather than diesel falling. He'd rather own the cheap end of that convergence.
Grains & agriculture Neutral
The world has spent fifteen to twenty years getting exceptional harvests just as demand for grain rose (as emerging-market diets added more meat, which takes a lot of grain to produce). Rozencwajg's warning is about complacency: everyone now assumes those yields continue — "it's a little bit like a trade where everyone's on one side of the boat." The question isn't whether a disaster is forecast, it's how much damage even a small disruption would do to a crop that is priced for perfection.
He deliberately does not call the direction. A very strong El Niño historically brings more volatile weather, which hurts yields, but also wetter summers — which would actually help a US farm belt currently suffering drought and depleted groundwater, and would be bearish for crop prices. Offsetting that, the drought is worldwide (Europe, the Black Sea, Brazil), a failed Indian monsoon would be serious, and farmers didn't get enough fertilizer onto the fields this year. His stance is a tail risk to monitor, not a position to hold.
US refiners Negative
US refiners are the businesses earning those record fuel-processing margins, so the obvious way to play a diesel shortage would be to buy them. G&R deliberately didn't: "we didn't invest in the US refiners… we thought that the crude oil molecule was the mispriced asset."
The logic is about where the problem started. Refining margins are wide because 10 million barrels a day of oil production was switched off upstream and refineries elsewhere in the world went dark — a symptom, not the source. Once the shortage is resolved the way he expects (crude rising to meet fuel prices rather than fuel prices falling), refiners lose: they'd be paying more for crude while their selling price stays put. His implied bet is exactly that — "the crack spread would come down with oil benefiting."
Summary & timestamps derived from the public YouTube video (transcript in transcript.html) for personal study. Not investment advice. © Peak Prosperity / Adam Rozencwajg & Goehring & Rozencwajg for source material.