Six checks. Two of them - the curve and what the product actually holds - decide more of a five-year outcome than any view about supply and demand, and both have exact published answers.
13 min read · Level 3
Most investment appraisal has an uncomfortable property: the things that decide the outcome are exactly the things nobody can know, and the work consists of making a forecast and being honest about its uncertainty. This class is a rare exception, and I think the exception is underexploited to a degree that is almost strange. Two of the six checks here - the shape of the curve and what a product actually holds - have exact answers, published today, free, taking about two minutes between them, and in the illustration in this lesson they were worth 24.99% of a five-year commitment. That is larger than most of what a confident and correct view about global supply would have delivered, and it required no view at all. The reason it is skipped is, I suspect, that it does not feel like investing. Reading a factsheet and a curve is administration, and administration has no narrative, whereas forming a view about electrification or monetary conditions is interesting and can be discussed. So attention flows to the part that is enjoyable and unknowable, and away from the part that is dull and exact. If I had one rule for this asset class it would be an ordering rule rather than an analytical one: do the two exact checks first, write the annual percentage they produce at the top of the page, and only then decide whether you have a view worth putting that much money behind. Quite often the exact checks will have already answered the question.
Six checks on one specific commodity exposure, ordered by how much each one moves a five-year outcome.
One. What actually drives demand for this good. Commodities are not one thing. An industrial metal is consumed by manufacturing and its demand follows construction, vehicles and electrification. A precious metal is partly industrial and partly monetary, held by people who want something that is not a promise. An agricultural good is consumed and regrown every year. A fuel is burned. These have almost nothing in common except the word, and a driver that matters for one is noise for another. Write which kind you are looking at and name the two or three things that actually move its consumption.
Two. Where the supply comes from, and what it costs to produce. How concentrated is production by country and by producer, how long does new supply take to arrive, and what does the marginal producer spend to get a unit out of the ground. That last figure is the closest thing this class has to an anchor:
Spot price 2,000 All-in cost for a high-cost producer 1,450 The spot price is 37.9310% above it
It is not a floor. Producers can and do operate below cost for a while, and low-cost producers keep going long after high-cost ones stop. What it gives you is a sense of where supply starts to withdraw, which is more than anything else in this class offers.
Three. The curve, and what it costs you to hold the exposure through it. Lesson three, in a minute: spot, three, six and twelve months, converted to an annual percentage. At the carry illustrated there, a futures route costs 5.4094% a year and a physical route costs its storage.
Four. What the product you are buying actually holds. Allocated metal recorded against the product, an unallocated claim, futures contracts, or shares in producers - which is not this asset class at all but a set of equity holdings whose fortunes are related to it. Four different instruments, one shelf.
Five. The total cost of ownership over your intended period. Not the headline fee. Run the whole thing. Over five years with the spot price unchanged:
Physical, 4.0% premium, 0.65% carry, 2.0% selling discount (1 / 1.04) x 0.9935^5 x 0.98 = 91.2078% of the outlay a loss of 8.7922%
Listed physically backed, 0.40% a year, 0.05% spread each way 0.9995 x 0.996^5 x 0.9995 = 97.9179% a loss of 2.0821%
Futures-based, 0.75% a year, rolling into the contango above 0.9925^5 x (1 - 0.013807)^20 = 72.9279% a loss of 27.0721%
On a 50,000 commitment: 45,603.91, 48,958.97 and 36,463.93
Six. How you would sell, and what that costs. A listed product sells in seconds at a small spread. Physical metal is sold to a dealer at a discount you can quote today. A futures position is closed at the prevailing price. Establish all three before you need one of them.
Because the two checks that decide most of a five-year outcome are the two that require no judgement at all, and the check that requires the most judgement decides the least.
Look at the range in check five. On an unchanged spot price, the same commitment to the same metal over the same five years ends at 45,603.91, 48,958.97 or 36,463.93. The gap between the best and the worst is 12,495.04 on 50,000, which is 24.99% of the commitment - and not one currency unit of it came from a view about the metal. It came from the route.
Now compare that with what a view about the metal would have to achieve. To cover the difference between the futures route and the listed one, the spot price would have to rise by 34.27% - 0.979179 / 0.729279 - over five years, purely to arrive back where the cheaper route started. That is a large forecast being asked to repair a decision that could have been made for free.
This is the general shape of diligence in this class and it is unusual. In equities the analysis is the work: the numbers that matter are forecasts, the disagreement is about the future, and there is no way to remove the uncertainty. Here, two of the six checks - the curve and what the product holds - have exact, published, current answers, and between them they account for more of a five-year outcome than any opinion about supply.
That does not make checks one and two pointless. It makes them the right place for judgement rather than the only place for effort. And they are where the genuinely interesting work is, because commodity demand is connected to physical activity in a way that rewards understanding something concrete: a metal used in electrical wiring behaves differently from one held as a monetary asset, and somebody who has confused the two will misread every piece of news about either.
Check two deserves one more sentence, because the marginal cost of production is the closest thing this class has to a valuation anchor and it is routinely overstated. A spot price 37.9310% above a high-cost producer's all-in cost does not mean the price cannot fall below 1,450 - it means that if it does, supply from that producer eventually stops, and eventually can be a long time. Anybody treating it as a floor has converted a soft supply response into a hard guarantee. It is a reference point in a class that otherwise has none, and that is worth a great deal without being worth what people ask of it.
And check six matters more here than it sounds, because the routes differ so much. A holder who bought physical metal because it is a thing that owes nothing to anybody has also bought an exit that requires finding a dealer, agreeing a discount and moving a physical object. That is entirely manageable and it is not a button.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
Two people who agreed that Metal M was worth owning and disagreed about nothing else.
Both commit 50,000 for five years. The spot price begins at 2,000 and ends at 2,000. The curve is in contango at the rate lesson three illustrates throughout.
One bought the futures-based product, because its annual fee of 0.75% looked comparable to everything else and the platform offered it first.
0.9925^5 x (1 - 0.013807)^20 = 72.9279%, so 36,463.93
One bought the listed physically backed product at 0.40% with a 0.05% spread each way.
0.9995 x 0.996^5 x 0.9995 = 97.9179%, so 48,958.97
A difference of 12,495.04, which is 24.99% of the commitment, on an identical view of an identical metal over an identical five years. The first investor did nothing wrong in any ordinary sense: the product was honest, it tracked its index, and its fee really was 0.75%. What they did not do was read one published curve and one line of documentation, and that omission was worth more than a correct forecast of the metal would have been.
Now suppose both had been right about the price as well, and spot rose 30.00% over the five years:
Futures-based 1.30 x 0.729279 = 94.8063% of the outlay still a loss of 5.1937% Listed physical 1.30 x 0.979179 = 127.2933% a gain of 27.2933%
The first investor was right about the metal, held for the full period, and finished below where they started. That is the shape this class produces when the route is chosen carelessly, and it is why check five belongs in the appraisal rather than in a footnote.
What the example is not is an argument that futures-based products are bad. In backwardation the arithmetic reverses and that route is the cheaper one. The argument is that the curve was published, the documentation was one line, and both were available before the money moved.
Treating all commodities as one asset class. An industrial metal, a precious metal, an agricultural good and a fuel share a word and very little else. A demand driver that matters enormously for one is noise for another, and somebody who has not written down which kind they hold will misread every piece of news about it.
Reading the marginal cost of production as a price floor. A spot price 37.9310% above a high-cost producer's all-in cost says where supply starts to withdraw, slowly, not where the price stops. Producers operate below cost for extended periods, and low-cost producers continue long after high-cost ones stop.
Comparing products on the headline fee. Over five years with an unchanged spot price, a 0.75% futures-based product returned 72.9279% of the outlay and a 0.40% physically backed one returned 97.9179%. The fee ranked them the same way; the outcome differed by 24.99% of the commitment.
Buying a producer's shares as a commodity exposure without saying so. Shares in miners or growers are equities, with management, debt, cost inflation and country risk of their own. They are related to the commodity and they are not it, and every check in public-equities-evaluating-one applies to them instead of these.
Checking the exit after buying. A listed product sells in seconds; physical metal requires a dealer, a discount and the movement of an object. Both are manageable and only one is a button, and the difference is worth knowing before it is needed.
Doing the hard check and skipping the easy ones. The supply-and-demand work is where the judgement is, and the curve and the holdings are where most of a five-year outcome is. Both matter; only one of them has an exact answer available today for free.
Industrial, precious, agricultural or energy, and then the two or three concrete things that change how much of it gets used. This determines which news is relevant to you and which is noise, and getting it wrong means misreading everything that follows.
Concentration by country and by producer, how long new supply takes to arrive, and an all-in cost figure for a marginal producer. Against a spot of 2,000 an all-in cost of 1,450 puts the price 37.9310% above it. Record it as a reference point, and write beside it that it is not a floor.
Spot, three, six and twelve months, published and current. At the carry in lesson three that is 5.6% a year, and rolling a three-month contract quarterly costs 5.4094% over the year. One minute, exact answer, and it is one of the two largest numbers in the appraisal.
Allocated metal, an unallocated claim, futures contracts, or producer shares. Four different instruments sit on one shelf under similar names, and only the documentation distinguishes them. Write the sentence down with the page you took it from.
Entry cost, annual carry or fee, roll if any, exit cost, compounded across the years you intend to hold. Over five years with an unchanged spot price that gave 91.2078%, 97.9179% and 72.9279% of the outlay - on 50,000, a spread of 12,495.04 between best and worst.
Get a dealer's current buying price for the physical route rather than assuming the spot price, check the spread on the listed one, and know how a futures position is closed. This is the check people leave until they need it, which is the one moment it cannot be done calmly.
On any specific commodity exposure before committing, when choosing between two products that name the same commodity, and when re-appraising a holding after the curve has changed shape.
Do not use this lesson to decide how much of this belongs in a portfolio - that is lesson six. And do not carry the cost figures here to another commodity; storage, production costs and curve shapes differ enormously between them.
Search a broker platform for a commodity by name and the results routinely mix physically backed products, futures-based products, leveraged products and funds holding producer shares, under names that differ by a word or two. All of them will rise when the commodity rises, which is enough to make them look interchangeable in any short comparison, and their behaviour diverges over longer periods for reasons that have nothing to do with the commodity: a roll cost, a producer's debt load, a leveraged product's daily reset. Buyers commonly select on the first two columns they are shown, which are usually the name and the annual fee, and discover the difference years later when two products that tracked the same word have returned very different amounts.
Never select a commodity product from a search result. Open the documentation, find the sentence that says what it holds, and select on that.
Create a free account to read the rest and apply it to your own investing.