Actionable insights — Commodity Weakness Is Coming, Here's What to Buy

58:08 1. Decide beta or alpha before picking a name — and admit most people should stop at beta

1. Define the two returns separately. Beta is what you earn simply from owning the sector while the sector outperforms — "beta defined as the extent to which that industry outperforms the market as a whole." Alpha is the extra you earn from picking the right company inside it.
2. Ask whether the beta alone clears your objective over your horizon. His answer for resources: "by 2030, 2031, you will have made enough buying the beta that you don't need to buy the alpha."
3. Qualify yourself, not the stock. The entry ticket to alpha is work you will actually do — "most of your listeners, like most of the population, aren't prepared to do the work that would allow them to buy smaller, more speculative issues" — plus, for a speculation, money whose loss "doesn't jeopardize their child's college education."
4. If beta is enough, buy the best of the beta rather than the cheapest: a long capital-allocation record plus, in a shortage thesis, a company that is actually spending sustaining capital.
5. Deliberately ladder any list you publish — "I tried to select three stocks that had varying degrees of risk, varying degrees of reward and would appeal to the different constituencies" — so a reader can find their own rung instead of buying the whole list.

• Your own hours per month. The alpha bucket's entry requirement is work, and it does not shrink when the tape is easy.
• A recommendation with no stated buyer — it is either wrong for you or accidentally right.
• The horizon on which "the beta is enough" is claimed. His is 2030-31; a two-year holder gets a different answer.

47:35 2. Juxtapose the nature of the risk against the size of the prize — every jurisdiction is risky

1. Start from the premise that eliminates the false safe harbour: "what I have learned is that every jurisdiction in the world is risky." There is no risk-free address, only risks you have stopped noticing.
2. Name the specific mechanism by which you would be harmed in this jurisdiction, in one sentence. Poland: "the political risk in Poland isn't being shot. It's the Polish government taking 80% of your profits by way of tax." Sinaloa: violence. California: "nobody shot me, but they delayed my permitting by 13 years."
3. Note that the English-language, legislative form of the risk is not automatically the milder one: "political risk can be Caucasian speaking English through the legislature," and his personal worst case came from a US state, not a failed one.
4. Set the named risk against the specific prize, not against a general return: "what you need to juxtapose is the nature of the risk relative to the size of the prize."
5. Ask whether the prize is substitutable. If an equivalent asset exists in a calmer place, the risk is not worth taking; if it does not, the calculation changes: "it's very difficult to find a deposit of this caliber elsewhere… Would I prefer that this deposit be in Nevada? Yes, ma'am. But it isn't."
6. Price your own capacity to carry that particular risk, and say so out loud when it exceeds a normal investor's: "highly risky. People who can't afford to take speculative risks should look elsewhere."

• Whether the risk you are underwriting is bounded (a tax rate, a permitting delay) or unbounded (violence). The first has a number; the second does not.
• Substitutability: how many assets of this quality exist elsewhere, and what they cost.
• Your own residency bias. He treats the People's Republic of California and the Ayatollahs in Alberta as the same category of finding.

12:43 3. Buy the thing you were going to buy anyway when it goes on sale — the physical-goods test

1. Separate the two questions people habitually merge: do I want to own this over my horizon? and where is the price going this quarter? A bearish answer to the second is only a problem if the answer to the first is no.
2. Establish that you are a net buyer, not a net seller, of the asset. "Although I own a fair bit of gold, I'd like to own a lot more."
3. If you are a buyer, treat a falling price as the intended outcome rather than an adverse one: "if they're looking to buy those items, they should be happy that prices are going down."
4. Apply the everyday behaviour as the check — "when you go shopping for clothes, you shop for sales" — and notice where you do the opposite: "when people buy financial goods, they seem to want to pay more. When people buy physical goods, they seem to want to pay less."
5. Convert the bearish view into a shopping list rather than a defensive posture: "what's going to happen in the balance of 2026 is that goods that I want to buy go on sale. Pretty obviously I'm going to buy them."

• The tell that you are not really a buyer: if a lower price makes you anxious rather than acquisitive, your stated horizon is not your real one.
• A price fall driven by something that also breaks the thesis — that is a different event from a sale.
• Whether you have the cash to act. A shopping list without funding is a forecast, not a plan.

20:12 4. Ask whether a commodity shortage is artificial or structural — one can be undone by an announcement

1. Classify the cause. An artificial shortage comes from politics and war and "doesn't have anything to do with production difficulties" — the barrels exist, they are blocked. A structural shortage comes from capacity that was never built.
2. Ask the single diagnostic question: could one event end this? "A shortage that could be solved by an armistice" is artificial by definition; the other kind, "we won't be able to end that shortage with an armistice."
3. Size the buffers absorbing the artificial shortage, because they determine how long it stays cheap-looking: demand destruction at the poor end of the market, floating inventory, strategic reserves ("the US strategic petroleum reserve, which is being drained at a very rapid rate," plus China and Japan), swing production, and substitution (LNG for oil).
4. Track those buffers depleting. When they are gone the price stops pricing an anticipated shortage and starts pricing "an actual shortage" — a different regime, "and you'll see something very different."
5. Date the structural shortage from the spending gap, not from sentiment: more than "a billion dollars a day" of missing sustaining capital, "cumulative and compounding," gives "by 2029, 2030, 2031… a structural shortage of oil."
6. Position in the structural one and tolerate the artificial one collapsing on you: "as a consequence of my belief in oil prices in 2030, I'm willing to take the near-term risk."

• Strategic-reserve and floating-storage levels — the countdown clock on the artificial shortage.
• Industry sustaining-capital spend versus depletion, which is the only thing that fixes the structural one.
• Forecasts being used to justify the underspend. He treats the 2030 peak-demand calls as the cause, not a coincidence: "they were wrong. They were simply wrong."

15:32 5. Price gold off negative real rates, not off the conflict headlines

1. Discard the intuitive driver first, with the evidence: "in 50 years of studying the gold price, I have learned that gold is remarkably resilient to conflict." A war can raise commodity prices generally and still not move gold.
2. Substitute the two he does use: "deteriorating faith in the purchasing power of the medium of exchange and negative real interest rates."
3. Compute the real rate yourself. Take the yield on "commonly held savings products like the US 10-year Treasury," subtract your estimate of purchasing-power decay rather than the published index. His: 8-10%. "You aren't getting a 4.6% yield, you're losing 2.4 or 3.4 or 4.4."
4. Treat a negative result as the whole thesis: when the safe store of value is guaranteed to shrink, gold's lack of a yield stops being a cost.
5. Separate the near-term direction from the driver. Rising nominal rates strengthen the dollar and pressure gold in the short run even while real rates stay negative — which is exactly the softness he is buying into.
6. Express the long-term target as a mirror of the currency, not as a valuation: a dollar losing "as much as 75%" of its purchasing power in ten years implies a nominal gold price that rises "in a way that mirrors the decline."

• The real yield crossing back into positive territory on your inflation number — that, not a headline, ends the thesis.
• Your own inflation estimate being unexamined; the entire calculation lives on that one input.
• Commentary attributing gold moves to geopolitics. On his data that correlation is weak, and mistaking it means selling on peace.

39:40 6. Read takeover inevitability from cost of capital and passive flows, not from rumour

1. Stop treating a bid as a lottery ticket and start treating it as arithmetic: "I regard it as icing on the cake, but I also regard it as inevitable."
2. Trace the advantage of scale in order: "bigger companies with bigger asset bases have larger trading volumes, attract more passive and ETF flows. The truth is that bigger companies just because they're bigger enjoy a lower cost of capital."
3. Conclude that a larger buyer can rationally pay more for the same asset than its current owner's market values it at — which is what makes the transaction happen rather than merely being desirable.
4. Screen candidates on the two attributes an acquirer is actually buying: capital efficiency and asset breadth. "Athabasca is a prime target… because of its capital efficiency and its very very very broad asset base."
5. Look for the prior instance in the same peer group — the pattern is usually already visible: "my favorite Canadian oil equity for quite some time was ARC… Shell noticed the same thing."
6. Apply the same mechanism to mergers, where it acts on the surviving company: combining lifts the market cap over index thresholds, so "as a consequence of larger market capitalizations, they now get a lot more index buying."

• Index-inclusion thresholds and average daily traded volume — the mechanical bid that follows scale.
• A valuation gap between small and large peers that persists; it resolves either by arbitrage or by acquisition.
• The other side of the trade: after a deal closes, the arbitrage holders sell. That is a price fall with no business cause — see insight 8.

50:59 7. Test social licence on local capacity, not on local sentiment

1. Separate genuine opposition from manufactured opposition and say which you have found: "there are real social and political challenges that they face, not fake ones — local opposition."
2. Then ask the sharper question — can the community that must decide actually evaluate what it is being offered? The failure mode he names is not hostility: "there is not sufficient local capacity to govern this deposit that needs to be developed."
3. Price the consequence honestly: "you run the risk that a population that doesn't have sufficient capacity will make a poor decision" — including one against its own interest.
4. Require the buyer to do the fieldwork rather than trusting the company's telling: "you need to acquaint yourself with that."
5. Invert it into an opportunity screen: a business whose model builds that missing capacity acquires a structural advantage nobody can outbid.

• Whether a community has independent technical advice, or only the company's presentation.
• Permitting timelines lengthening without a formal refusal — capacity failure usually looks like delay, not rejection.
• Companies whose competitive advantage is relational rather than financial; those moats do not show up in a screen.

1:03:29 8. When a stock is weak, identify who is selling before deciding whether anything is wrong

1. Ask what the sellers originally bought the stock for, not what the business is doing now.
2. Look for a reason that has expired rather than a thesis that has broken: holders who "owned those stocks for a takeover" lost their reason the moment the takeover completed, so "they sold the stock."
3. Label the condition as temporary and finite: "this is a hiatus."
4. Name the buyer who has to replace them, because that is the actual catalyst: someone "who understand[s] that the value of the combination is worth more than the price of the constituent parts."
5. Confirm the operating story is intact and improving independently — nameplate capacity reached, capacity expansions, acquired development assets flowing to the income statement.
6. Only then buy the weakness, and be explicit that the weakness is the reason: "and that's why I did it."

• Post-deal share registers — arbitrage funds show up in filings and their exit is predictable and finite.
• Whether operating milestones keep landing while the price falls; that divergence is the signature of a flow-driven decline.
• The reverse case: a price rising because holders bought for a reason that is about to expire.

1:05:58 9. Screen for "adult supervision" — an owner big enough to stop management

1. Look past the executive team to the share register: is there a family or strategic holder with a stake large enough that their own money disciplines the decisions? "20% of the company's owned by the Lundin family… once again, I have adult supervision."
2. Prefer alignment created by ownership over alignment promised by incentive plans: "I like management that's shareholder-centric because they're big shareholders," and distributions continue "because they're the largest shareholders."
3. Count a strategic partner as a second form of it — a major that has looked at the geology and stayed (Rio Tinto at Tamarack), or a merchant bank supplying "adult supervision" during the exploration-to-production transition.
4. Weight the family's duration, not its name: forty years with the Lundins, four decades with Ross Beaty, a whole adult life with the Goodmans, four decades and three CEOs at Agnico.
5. Do not let the supervision override the business question. Where a supervised company changes what it does, he withholds judgement rather than extending the credit: "I reserve my comment on that until I see how they proposed to effect that strategy."

• Insider ownership expressed in dollars at risk, not in percentages of comp.
• A controlling holder whose interests diverge from minorities' — supervision is only useful when they hold the same shares you do.
• A supervised company changing business model; the track record does not transfer across the change.

1:01:31 10. Rank peers on a 30-year capital-allocation record, then look for the unearned geographic premium

1. Pick the one metric that decides survival in a self-consuming industry — how well management has bought, built and replaced the asset base — and measure it over a period long enough to include a full cycle: "the capital allocation decisions among Barrick, Newmont, and Agnico over the last 30 years."
2. Rank rather than score, and accept the answer even when it is lopsided: "first, second, and third all belong to Agnico. They've done such a better job that there's no comparison."
3. Check for countercyclical behaviour as the fingerprint of a good allocator: "they continue to acquire and explore during the bad period. Now that the good period is here, they don't need to make overpriced acquisitions."
4. Include the soft variables that have hard costs. Turnover at "a third of the level" of peers reduces recruiting, training and injury expense — an operating-cost edge that never appears in a resource statement.
5. Then look for the premium the market is paying for something other than performance. His: "investors like Canadian investors are ethnocentric and so you get a tremendous bonus for being involved in northern Nevada in the Carlin," while the Abitibi — no worse "geologically or in terms of infrastructure" — does not get it.
6. Buy the best record trading without the premium: "unfounded, particularly unfounded given the returns on capital employed and returns on invested capital."

• Reserve replacement over a decade, not production growth over a quarter — the acquisition that flatters this year's output can destroy the record.
• Valuation premiums that track address, index membership or familiarity rather than returns on capital.
• Employee turnover and safety statistics as leading indicators of cost discipline.

29:55 11. Publish nothing where you have no edge — including a defence of something you own

1. Split any question into the parts you can judge and the parts you cannot, and answer only the first. On Abaxx he rates the founder ("a very formidable competitor who I like") and the mission ("we need more competition, technologically driven competition, in the exchange business") and stops.
2. State the boundary in terms of your own record rather than as modesty: "my track record in science and technology is unblemished by success" — evidenced by having tried in his thirties and being "amazed at my lack of success."
3. Apply the boundary symmetrically. It stops you rating and stops you defending: "I did not take on the short report because I don't have a high regard for my own ability to analyze the technology."
4. Contrast against a case where you did have the edge, so the rule is visible as a rule: he rebutted a short attack on a silver company "where I had the advantage of understanding a lot more about the silver mining business than I do about exchange technology."
5. Route the question to someone competent instead of guessing — offer to interview the founder so a "financially sophisticated, rational observer" can judge the rebuttal.
6. Build the boundary into your time budget, not just your mouth: "if you ask me a question about laws as an example, I don't know and I would tell you that and I don't waste any time on it whatsoever."

• Yourself defending a holding on grounds you could not have used to buy it. That is the boundary being crossed.
• An analyst whose confidence is constant across subjects — the real ones have visible edges.
• Where a specialist's team supplies the edge instead: "I employ geologists, I employ engineers, I employ financial analysts. So I don't do it all myself."