Actionable insights — Reading a Commodity Through Inventories, Not Price

1. When the price is calm, go and measure the inventory — the physical market leads the quote

1. Stop treating the front-month price as the state of the market. It is a clearing price for the marginal barrel today, and it can stay quiet while the buffer behind it is consumed.
2. Measure the buffer instead: days of cover, absolute stock levels, and — the decisive one — proximity to tank bottoms, the operationally unusable minimum below which inventory cannot fall.
3. Check where the drawdown is happening in the barrel. Stress often appears first in a product (diesel, jet) rather than in crude, because refining is the constrained step.
4. Corroborate with the behaviour of physical buyers, not with sentiment surveys: are consumers of the commodity bidding for prompt cargoes, paying up for delivery, panicking?
5. Treat "inventory near tank bottom with a continuing draw" as an unstable state — the price has no shock absorber left.

• Weekly distillate stocks vs the five-year range and vs estimated minimum operating levels; days of cover; physical differentials and prompt-vs-deferred spreads (backwardation steepening); refinery utilisation and outage tallies in Russia/China/Middle East.

2. Ask what happens on the resolution — restocking is demand stacked on consumption

1. For any supply disruption, model two demand curves: ongoing consumption, and the one-off refill of every inventory drawn down while the disruption lasted.
2. Note that the market prices the end of the disruption as bearish, because it restores supply — and usually ignores the refill, which arrives at the same moment and is concentrated in time.
3. Size the refill: cumulative draw to date is the minimum quantity that must be re-bought. The longer the disruption ran, the larger the bullish event hiding inside its resolution.
4. This inverts the obvious trade: peace can be the buy signal, not the sell signal, when the buffer has been exhausted.

• Cumulative inventory draw since the disruption began (the refill quantity); tanker fixtures and freight rates on reopening; strategic-reserve replenishment mandates; the shape of the curve flipping as prompt demand returns.

3. Test the supply answer before believing the demand story — find the marginal barrel and check whether it can still grow

1. Identify which single source has supplied most of the growth in the last cycle. Concentration of growth is a fragility, not a strength.
2. Test that source's ability to keep growing from its inputs, not its output: capital spending, rig counts, inventory of drillable locations, well productivity trends.
3. If investment has fallen, project the point at which growth turns negative — decline rates make this arithmetic, not opinion, on a short lag.
4. Then ask the elimination question: is there any other project or region large enough to replace it on that timeline? If not, the pricing power moves to whoever is left.
5. Look for the historical analogue in which the same source rolled over, and use it to bound the magnitude rather than to predict the path.

• Monthly U.S. crude production prints (EIA 914 revisions especially) rolling from positive to negative y/y; shale capex guidance and rig counts; tier-1 inventory-life disclosures; OPEC spare-capacity estimates shrinking.

4. Express a long-horizon commodity view in equities, not futures

1. Ask over what horizon the thesis pays. Multi-year theses die in futures because roll cost, margin calls and expiry force you to be right about timing as well as direction.
2. Use equities to hold the view through the flat period: an operating company generates cash while you wait and re-rates when the price moves.
3. Accept the trade-off explicitly — equities add company risk (balance sheet, management, cost inflation) in exchange for removing time risk.
4. Where the market has become bearish on the commodity, prefer equity that is priced for the low commodity price: you are then paid twice, by the price and by the multiple.

• The commodity curve's shape (a steep contango is a direct tax on futures expression); implied commodity price in equity valuations vs spot; the equity's break-even price deck.

5. Choose the instrument with the fewest offsetting exposures — diversification inside a company dilutes the call

1. Write the thesis as a single variable ("crude above $100 for a sustained period").
2. For each candidate, decompose earnings by segment and ask which segments are negatively or non-correlated to that variable. Those segments are hedges you did not want.
3. Compute a rough sensitivity: what fraction of gross profit actually moves with the variable? Rank candidates by that fraction, not by quality or size.
4. Deliberately reject the household name if its diversification is what makes it safe — safety here is the same thing as muted payoff.
5. Then, separately, decide how much of the resulting volatility to hold, at the portfolio level rather than inside the security.

• Segment disclosure in the 10-K (what share of EBIT is upstream vs downstream vs chemicals); realised price sensitivity per $10/bbl in company guidance; hedging programs, which silently remove the exposure you are paying for.

6. Match asset life to forecast horizon — for a multi-year price call, buy the slow-decline barrel

1. Write down the years over which the thesis pays. Then require the asset to still be producing across all of them without new capital.
2. Rank producers by decline rate and reserve life, not by growth rate. Fast-decline assets convert a price forecast into a single good year plus a treadmill of reinvestment.
3. Check maintenance capital: how much spending is needed just to hold production flat? Free cash flow at the forecast price is what you are buying, and the treadmill eats it.
4. Prefer assets whose scarcity increases as the fast-decline competitor rolls over — the same event that drives the price also raises the relative value of longevity.

• Corporate decline rates and reserve-life indices; sustaining-capital-per-barrel disclosures; free cash flow break-even at a stated price deck; whether shale peers' guidance shifts from growth to maintenance.

7. Buy the service capacity that the last downturn destroyed, once the demand for it must return

1. When the marginal supply source must shift (onshore to offshore, one basin to another), find the service capacity that shift requires.
2. Check what that capacity did in the prior downturn: bankruptcies, scrapping and restructuring shrink the fleet permanently and leave survivors with clean balance sheets.
3. Confirm demand is actually turning — contract awards, tender activity, operators publicly reconsidering the segment — before paying for the recovery.
4. Price it off utilisation and day rates rather than earnings: in a capacity-short services market, both move non-linearly once utilisation crosses the tight threshold.

• Offshore rig utilisation and day rates; contract backlog and tender counts; final investment decisions on deepwater projects; newbuild orders (the signal that the capacity shortage is being cured, i.e. the exit).

8. Hold a latent-shock thesis with an explicit rule for being early

1. Recognise the signature of a latent shock: a physical constraint is building, prices are calm, and the people warning look like alarmists. Nothing about the calm falsifies the thesis.
2. Write the falsification test in physical terms up front — inventories rebuilding, supply growth resuming — so "the price hasn't moved" cannot be mistaken for evidence against you.
3. Structure the position to survive the flat period (equities over futures, positive-cash-flow operators, no leverage that a drawdown can call).
4. Say out loud what you got wrong meanwhile; a manager who names their misses is running a testable process.
5. Expect the repricing to be discontinuous — you will not be able to add on the way up.

• The physical falsifiers specifically: distillate stocks rebuilding for several consecutive weeks, U.S. production growth re-accelerating, Hormuz traffic normalising with no import surge behind it. Any of those, not the price, is what should close the position.