@sdelsad/commodity-desk-daily 1.0.33 → 1.0.35

package/ep11.md

@@ -1,344 +0,0 @@
-# Storage, Elevation and Trade Flows
-
-## Market pulse
-
-**Pro Farmer walked out of the fields with seven and a half bushels less corn than USDA, and December corn went to a two-and-a-half-year high.**
-
-| Commodity | Contract | Price | Change | Change on week |
-|---|---|---|---|---|
-| Corn | Dec (CBOT) | 508½ c/bu | +5¢ | +25¼¢ |
-| Soybeans | Nov (CBOT) | 1239½ c/bu | +3¢ | +47¢ |
-| Soymeal | Sep (CBOT) | $317.70/st | +$2.00 | +$7.50 |
-| Soyoil | Sep (CBOT) | 69.35 c/lb | −183 pts | −9 pts |
-| Wheat SRW | Sep (CBOT) | 681½ c/bu | −1¼¢ | +6¾¢ |
-| Wheat HRW | Sep (KC) | 756¼ c/bu | −6¢ | +2¢ |
-| Wheat HRS | Sep (MGE) | 698¼ c/bu | −2½¢ | +20¢ |
-
-Pro Farmer closed its four-day tour on Friday evening with a national corn yield of **173.2 bu/ac** and a crop of 15.344 bn bu. USDA said 180.7 three weeks ago. That gap — more than four percent of the crop — is what carried December corn through five dollars and to its highest close since early 2024. Soybeans went the other way: Pro Farmer has 53.3 bu/ac against USDA's 52.7, a slightly *bigger* bean crop, yet November beans still added 47 cents on the week on flooding in the eastern belt, a weaker dollar and a solid week of export sales.
-
-Wheat was the quiet one. All three exchanges gave a little back on Friday after a firm week, with Minneapolis spring wheat the best of them at +20 cents.
-
-The number worth carrying into this week is not the yield. It is the crop. Even on Pro Farmer's reduced figure, the United States is about to harvest more than fifteen billion bushels of corn into a storage system that stopped being built six years ago.
-
-**The geopolitical read: when the export door shuts, the grain does not vanish — it queues.** More than 90% of Russia's grain export capacity in the Azov–Black Sea basin is currently offline. Three Novorossiysk terminals suspended last week, Taman stopped in late July, and navigation in the Sea of Azov has been suspended since July, leaving one working deepwater grain terminal in a basin that moved 46.3 mt last season. The transmission this time is not freight and it is not war-risk premium. It is storage. Russia has harvested about 140 mt, exporters have stopped buying because they cannot ship, and the grain is backing up inland. Fourth-class Russian wheat is around **12,000 roubles/t**, against 15,000 a year ago. The world price of wheat is rising and the Russian farmer's price is falling, in the same crop, at the same time. What sits between those two prices is storage and the ability to move.
-
-```chart
-{"type":"bar","unit":"cents per bushel","title":"Corn and beans took the week",
- "x":["Corn Dec","Beans Nov","SRW Sep","HRW Sep","HRS Sep"],
- "series":[{"name":"Change on the week","values":[25.25,47,6.75,2,20]}],
- "caption":"Beans added the most cents, but corn added the most meaning: a 25-cent week that ended at a two-and-a-half-year high, on a tour result four percent under USDA. Wheat went along for the ride.",
- "source":"CBOT, KC and Minneapolis settlements, week ending Friday 21 August 2026."}
-```
-
-## Key takeaways
-
-- An elevator is not long grain and it is not a speculator. It buys at the posted bid and sells the board within minutes. What it owns is space and the right to move grain through it.
-- A posted bid can be a refusal. When space is the binding constraint, the bid stops being a price and becomes a queue-management tool.
-- Measure both ends of a storage trade against **one** futures month and the arithmetic collapses to a single number: the basis you bought, against the basis you sold.
-- The carry in the curve can only be collected by somebody who has a bin. Without space, contango is a number on a screen that someone else will earn.
-- US storage capacity has been flat at roughly 25.3 bn bu since 2019, against a twenty-year trend that would have put it at 27.5 bn. That missing 2.2 bn bushels of shed is why harvest basis behaves the way it does.
-- On-farm storage ran at a record 80% of capacity on 1 December 2025, and system-wide surplus capacity was about 5% against a century average of 15% — the tightest since 1988.
-- Blending is the cheapest form change in the business: no chemistry, no plant, just a weighted average and two tanks.
-- The discount lot is cheap because most buyers physically cannot use it, which is why the blender usually sets the discount rather than taking it.
-- Protein, moisture and test weight average. Aflatoxin, live infestation and an unapproved genetic event do not — those are pass/fail on the whole lot, and pouring contaminated grain into clean grain gives you a bigger contaminated lot.
-- Replacement value — what it would cost to buy back right now what you just sold — is the only question that tells you whether you can do the trade again tomorrow.
-- Whoever controls the bottleneck sets replacement value. In a congested port in harvest week, no view on flat price competes with owning the space.
-
-## Vocabulary
-
-| Term | Meaning |
-|---|---|
-| **Country elevator** | The first commercial storage point off the farm, buying from growers and shipping onward by truck, rail or barge |
-| **Terminal elevator** | Large storage at a port, river or rail hub, whose business is blending, load-out speed and access rather than farm origination |
-| **Elevation margin** | The toll an elevator earns for taking grain in, conditioning it and loading it out, separate from any gain on the basis |
-| **Throughput** | The volume moved through a facility in a period, the number that actually pays for a fixed asset — capacity earns nothing standing still |
-| **Receiving capacity** | How fast an elevator can take grain in, in bushels or tonnes per hour, which is a different constraint from how much it can hold |
-| **Load-out capacity** | How fast an elevator can ship grain out, and the lever that decides whether a full house is a crisis or a rotation |
-| **Storage tariff** | The published charge for commercial storage, quoted in cents per bushel per month or per day, or in dollars per tonne per month |
-| **Shrink** | Weight lost when grain is dried to a safe keeping moisture, deducted as a percentage and a real cost to whoever owns the grain |
-| **Harvest basis** | The seasonal low in the cash-minus-futures spread, set when a year's crop arrives in six weeks into a pipe sized to move it over twelve months |
-| **Temporary storage** | Ground piles, bunkers and bags used when permanent capacity is full, cheap per bushel to build and expensive per bushel in spoilage and rehandling |
-| **Capacity utilisation** | The share of storage capacity actually occupied, and the single best leading indicator of what harvest basis is about to do |
-| **Blending** | Combining lots of different quality so the weighted average meets a contract specification, creating value from material nobody else can use |
-| **Segregation** | Keeping identities and grades physically apart in separate bins, the precondition for being able to blend deliberately later |
-| **Discount schedule** | The published table of price deductions for grain that falls outside a contract's grade limits, and the raw material of every blending trade |
-| **Pass/fail specification** | A contract term that cannot be met on average — contamination, infestation, unapproved events — where blending increases the affected tonnage instead of diluting it |
-| **Replacement value** | What it would cost to buy back today what you have just sold, the test of whether a price was genuinely good |
-| **Arb window** | The period during which a route's economics work, opening and shutting on freight, differentials and FX rather than on flat price |
-| **Bottleneck asset** | A facility with no near substitute at the moment it is needed, whose owner sets the price rather than quoting one |
-
-## Quiz
-
-**Q1.** A country elevator buys 3,000,000 bu of corn at harvest at 45 under the December, hedges by selling December futures, and plans to sell in late February at 15 under the March with March trading 18 cents over December. Compute the gross basis gain per bushel and in total. Now change two things: the Dec–March spread narrows to 6 cents by January, and the cash basis only recovers to 25 under the March. Recompute. State which of the two changes cost more, and explain why the elevator has meaningful influence over one of them and almost none over the other.
-
-**Q2.** A blender holds 45,000 t of milling wheat at 11.6% protein bought at $238/t, and 25,000 t at 13.0% bought at $259/t. He has a sale of 70,000 t against a 12.0% protein minimum at $251/t. Compute the blended protein, the weighted cost, and the margin per tonne and in total, before and after $3.00/t of elevation and handling. Then: the 45,000 t lot returns a falling number of 180 seconds against a contractual minimum of 250. Say whether the trade still works, and explain precisely why that test behaves differently from the protein test.
-
-**Q3.** A country elevator holds 3,000,000 bu of licensed space. It can receive 40,000 bu/hour and runs 12 hours a day for the 22 working days of harvest, over which its draw area will deliver 4,200,000 bu. Compute the receiving capacity for the period and say whether the binding constraint is throughput or space. Then compute the minimum daily load-out, in bushels a day, that would keep the house from filling. Finally, say what the posted bid has to do if that load-out is not achievable, and why that is not the same thing as the elevator having a bearish view on corn.
-
-**Q4.** *(Ep 10)* Ep 10 established that a Panamax was earning $18,964/day and that selling CFR with no vessel fixed leaves you structurally short freight. A desk instead buys FOB at origin, sells FOB at origin, and never touches the ocean — but the export berth is congested and its nominated vessel waits eight days beyond laytime on a 66,000 t cargo. Compute the demurrage-equivalent cost in dollars and in dollars per tonne at that day rate. Then explain why this is a storage problem wearing freight clothes, and which party carries it under FOB as against CFR.
-
-**Q5.** *(Ep 10)* Ep 10 said an FFA hedges a basket of named Baltic routes rather than your voyage. An elevator signs a twelve-month throughput agreement committing it to move 900,000 t through a port terminal it does not own. Explain why that exposure cannot be hedged with an FFA at all — be specific about what the exposure actually is — and name the instrument or arrangement that does cover it.
-
-**Q6.** *(Ep 8)* Using ep 8's board crush arithmetic, compute Friday's September board crush from meal at $317.70/short ton, oil at 69.35 c/lb and beans at $12.25/bu. Give the crush in dollars per bushel and the oil share as a percentage. Then explain how today's storage lesson changes where a crusher's real margin sits, given that the beans arrive in eight weeks and the plant runs for fifty-two.
-
-**Q7.** *(Ep 8)* Ep 8 argued that a wide crush is competed away through the bean basis at the gate rather than through new plant construction, because capacity takes two to three years to build. Explain why a crush plant sitting in a region with a storage deficit faces a structurally different bean basis in October than in June, say which direction that pushes its October run rate, and name the one thing that would reverse the conclusion.
-
-**Q8 — Conversion drill.** A US mill quotes long-grain rough rice at **$13.85 per cwt**. A competing Vietnamese cargo is offered at **$348.00 per tonne**. Put both on dollars per tonne, and say which is cheaper and by how much. Then: the terminal holding 8,400 t of that rough rice charges storage at **4.5 cents per cwt per month**. Convert that tariff to dollars per tonne per month and compute the monthly storage bill on the whole parcel.
-
-## SOLUTIONS (spoilers)
-
-**A1.** Put both ends against December and the arithmetic becomes one subtraction.
-
-| | Base case | Revised |
-|---|---|---|
-| Basis bought (vs Dec) | −45¢ | −45¢ |
-| Basis sold (vs Mar) | −15¢ | −25¢ |
-| Mar over Dec | +18¢ | +6¢ |
-| Basis sold, restated vs Dec | +3¢ | −19¢ |
-| **Gross gain** | **48¢** | **26¢** |
-| **On 3,000,000 bu** | **$1,440,000** | **$780,000** |
-
-A swing of **$660,000**, with the flat price hedged throughout.
-
-Splitting the damage: the spread narrowing from 18 to 6 costs 12 cents, and the weaker cash basis costs another 10. The **spread** cost more, and it is also the one the elevator cannot do anything about. The Dec–March spread is a market-wide statement about how much the market is willing to pay for storage, set by everybody's inventory and everybody's space. One elevator's decisions do not move it.
-
-The **cash basis** is the half the elevator genuinely works. It chooses who to bid, when to bid, which end-users and which shippers to court, whether to hold for a rail programme or a river bid, and whether to spend elevation margin to reach a better market. That is the business. The trap in the question is that the line an operator obsesses over — "am I getting my basis back?" — was the smaller of the two losses. The larger one arrived silently, on a screen, as the market told everyone at once that storage was worth less than it had been.
-
-**A2.** Protein first, because if it does not blend there is nothing to price.
-
-| Lot | Tonnes | Protein | Protein tonne-% | Cost $/t | Cost $ |
-|---|---|---|---|---|---|
-| A | 45,000 | 11.6% | 522,000 | 238.00 | 10,710,000 |
-| B | 25,000 | 13.0% | 325,000 | 259.00 | 6,475,000 |
-| **Blend** | **70,000** | **12.10%** | 847,000 | **245.50** | 17,185,000 |
-
-12.10% clears a 12.0% minimum with a tenth of a point to spare. The sale is at $251.00/t.
-
-| Line | $/t | Total |
-|---|---|---|
-| Sale | 251.00 | 17,570,000 |
-| Weighted cost | −245.50 | −17,185,000 |
-| **Gross margin** | **5.50** | **385,000** |
-| Elevation and handling | −3.00 | −210,000 |
-| **Net margin** | **2.50** | **175,000** |
-
-Now the falling number. **The trade does not work, and the arithmetic above is irrelevant.**
-
-Protein is a *concentration*. Mix two concentrations by weight and the result is the weighted mean, reliably and linearly. Falling number is not a concentration of anything — it is the Hagberg test's measure of how fast a flour-and-water slurry loses viscosity, which is a measure of **alpha-amylase enzyme activity** in sprout-damaged grain. Enzymes are catalysts. A small quantity of highly active material degrades starch far out of proportion to its weight share, and the test's response is strongly non-linear: a blend of a 180 lot and a 350 lot lands well below the weighted average, and the direction of the error is always against the seller.
-
-The practical rule on a desk is that falling number blends *down* much faster than it blends up, so a low-FN parcel is treated as contaminating rather than diluting. It also matters that ep 5 already priced this: a low falling number demotes milling wheat to feed at roughly $40/t. That is the real value of lot A, not a $13 protein discount — and at feed value the blend was never a $2.50/t margin. It was a $2.50/t margin sitting on top of an unrecognised $40/t writedown.
-
-**A3.** Receiving capacity for the harvest window:
-
-40,000 bu/hr × 12 hr × 22 days = **10,560,000 bu**
-
-Against 4,200,000 bu of deliveries, receiving is not remotely the constraint — the house could take the whole crop two and a half times over. **Space is the constraint.** 4.2 m bu of deliveries into 3.0 m bu of licensed space leaves 1,200,000 bu with nowhere to go.
-
-Minimum load-out to stay level: 1,200,000 ÷ 22 = **54,545 bu/day**, call it two and a half unit trains a month, or a barge every few days.
-
-If that load-out is not achievable — no rail cars allocated, river levels down, the terminal downstream full of its own crop — the posted bid has to widen until deliveries slow to what the house can absorb, or until the widened bid pays for ground piles and their rehandling. Either way the bid is doing physical work, not expressing a view.
-
-This is the distinction worth keeping. A bearish view says *corn is worth less than the board implies*. A full house says *corn delivered to me, this week, at this location, is worth less to me than corn delivered next month, because I have nowhere to put it.* The first is a statement about corn. The second is a statement about a building. They look identical on a bid sheet and they are completely different trades — which is exactly why a merchant who reads country bids as sentiment gets the direction of the next basis move wrong.
-
-**A4.** Eight days at $18,964:
-
-| | |
-|---|---|
-| Days over laytime | 8 |
-| Rate | $18,964/day |
-| **Demurrage-equivalent** | **$151,712** |
-| Cargo | 66,000 t |
-| **Per tonne** | **$2.30/t** |
-
-It is a storage problem in freight clothes because a berth is storage in motion. Congestion at the export elevator means grain cannot get out of the house and into the hold, so the queue that forms is a queue for *space and load-out capacity*, not for ocean tonnage. The vessel is simply the meter that is running while the problem is somewhere else — in the silo, on the conveyor, in the rail allocation.
-
-Who carries it depends on the term of sale, and this is the ep 4 point made concrete. Under **FOB**, the buyer charters, so the buyer's vessel is on demurrage — but the seller has undertaken to load within an agreed laytime, so the demurrage claim comes straight back at the seller under the sale contract. Under **CFR**, the seller charters and simply wears it directly against the shipowner. The economics are similar; the paperwork and the timing are not, and the difference is where the argument happens. Note also the asymmetry the question is really testing: a desk that never buys a tonne of freight can still lose $2.30/t to freight, because it sold an obligation to *load at a rate*, and loading rate is a property of a building it may not even own.
-
-**A5.** The throughput agreement is not a freight exposure at all, which is why no freight instrument touches it.
-
-What the elevator has sold is a **volume commitment**: 900,000 t must move through a terminal over twelve months. The risks attached to it are, in order: that the grain does not exist to move (a short crop in the draw area), that it exists but flows to a competing outlet (a rival bid, a rail programme, a better river basis), that the terminal cannot take it when it arrives (congestion, breakdown, a queue), and that the take-or-pay clause bites for the tonnes not moved. Those are origination, competition and access risks.
-
-An FFA settles against the average of a basket of Baltic route assessments over a calendar month. It pays out when *ocean freight rates* move. Ocean freight rates can be perfectly flat for the whole year while every one of the risks above destroys the contract — and conversely, freight can double without changing the elevator's obligation by a tonne. The correlation is not weak. It is close to structurally absent, and hedging with it would be adding a naked freight position to an unhedged throughput position.
-
-What actually covers it is **contractual, not financial**: a take-or-pay with a shortfall cap or a carry-forward of unused tonnes, a matching origination programme (forward or deferred-price purchases sized to the commitment), and a berth or window guarantee with the terminal operator giving priority access. Where a financial hedge helps at all, it is on the *basis* — the elevator can buy origin basis forward to lock the acquisition cost of the tonnes it has promised to move. The honest summary: an access risk is hedged by buying access.
-
-**A6.** The two multipliers, in the fixed order.
-
-| Leg | Price | Multiplier | $/bu |
-|---|---|---|---|
-| Meal | $317.70/short ton | × 0.022 | 6.9894 |
-| Oil | 69.35 c/lb | × 0.11 | 7.6285 |
-| **Gross product value** | | | **14.6179** |
-| Beans | $12.25/bu | | −12.2500 |
-| **Board crush** | | | **$2.368/bu** |
-
-Oil share = 7.6285 ÷ 14.6179 = **52.2%**. Oil is again more than half the value of the bushel.
-
-The storage point changes where that margin actually lives. The board crush is a *simultaneous* quote: three prices observed at the same instant. The plant's real problem is that its raw material arrives in an eight-week window and its output is sold across fifty-two weeks. So the crusher is, structurally, a storage business with a processing plant attached. It must either own the bins to carry beans from harvest through to summer, or buy them month by month from someone who does — and the price of doing that is the bean basis plus the carry in the bean curve.
-
-That means a crusher hedging only the board crush has hedged the smaller half of its problem. It has locked a margin per bushel and left open the question of whether it can source bushels at that basis for the other forty-four weeks. In a year where storage is scarce, the carry the crusher must pay to hold beans, or the basis premium it must pay to buy them later, comes straight out of a margin that the screen still shows as $2.37.
-
-**A7.** In **June** the bean is somebody's carried inventory. It has already been stored, financed and priced; the seller's alternative to selling is another month of storage cost, and the plant is competing against the *carry* in the curve. In **October** the bean is arriving on a truck from a field into a region with nowhere to put it. The plant's alternative bidder is a full elevator, and a full elevator bids to repel.
-
-So in October a crush plant in a storage-deficit region gets its beans **cheaper on the basis** — sometimes dramatically so — because it is one of the very few buyers in the county that can take delivery and immediately consume the grain rather than store it. A plant is a hole in the ground that never fills. That is a genuine competitive asset for six weeks a year.
-
-Direction: it pushes the October **run rate up**, toward capacity, and it makes the plant want to be running flat out precisely when the basis is at its seasonal low. This is the seasonal shape of crush margins that ep 8's board-crush arithmetic cannot see, because the board crush contains no basis.
-
-The one thing that reverses it: **meal**. Meal cannot be stored economically or for long — it degrades, it takes volume, and it has to be sold into a local feed market. If the plant runs flat out in October it must place October meal, and if the feed market in its own draw area is also flooded with a big local corn crop and cheap local feed alternatives, the meal basis collapses faster than the bean basis improves. At that point cheap beans buy you nothing, because the constraint has moved from the intake to the outlet — which is the same lesson as the elevator's, standing on its head.
-
-**A8 — Conversion drill.** One tonne is 2,204.62 lb, which is **22.0462 cwt**.
-
-| | Quote | Conversion | $/tonne |
-|---|---|---|---|
-| US rough rice | $13.85/cwt | × 22.0462 | **$305.34** |
-| Vietnamese cargo | $348.00/t | — | $348.00 |
-
-The US rice is cheaper by **$42.66/t**, or in the other currency of the trade, $348.00 ÷ 22.0462 = $15.79/cwt against $13.85 — a **$1.94/cwt** advantage. Both statements are the same fact; which one you say out loud depends on who you are talking to, and getting that wrong in front of a counterparty is how you sound like you have never traded rice.
-
-Now the storage tariff:
-
-4.5 c/cwt/month × 22.0462 = **$0.9921/t/month**, call it 99 cents.
-
-On 8,400 t: 8,400 × 0.9921 = **$8,333 per month**.
-
-Worth noticing how small that is relative to the parcel. The rice at $305.34/t is worth $2,564,856, so a month of storage is 0.32% of the value — while a month of interest at 5% is 0.42%, and larger. On a low-value, high-bulk commodity the tariff dominates; on rice it does not, and the financing does. That is why the store-or-sell answer is different for rice than for corn even when the curve looks the same shape.
-
-## The written edition
-
-Last year the United States built almost no new grain storage. This autumn it will harvest something over fifteen billion bushels of corn. Those two facts meet in September, and the place where they meet has a name. It is called the basis.
-
-### An elevator is not a warehouse
-
-The most common mistake about country elevators is to think of them as speculators with buildings, or as warehouses with a price list. They are neither.
-
-An elevator buys grain at the posted bid and sells futures against it within minutes. Flat price is dead on arrival — the same discipline the whole show has been describing since ep 2. What the elevator owns is **space**, and the right to move grain through that space.
-
-That produces two revenue lines, and they are genuinely separate businesses:
-
-| Line | What it is | What pays for it |
-|---|---|---|
-| **Elevation margin** | Taking grain in, drying and conditioning it, loading it out | Throughput — volume moved, regardless of price |
-| **Basis and carry** | Buying the basis cheap at harvest, selling it back later | Ownership of space when space is scarce |
-
-Two units of the trade are worth stating precisely, because they are where the money hides. **Commercial storage** is quoted in cents per bushel per month, sometimes per day, and outside the US in dollars per tonne per month. **Shrink** is quoted in percent: it is the weight that disappears when wet grain is dried down to a moisture at which it will keep. Neither of them is a price view. Both of them are deductions taken from someone, and the argument about which someone is most of what an origination contract is for.
-
-### The bid that is not a price
-
-Here is how a posted bid gets given in the third week of October.
-
-> **FARMER:** What are you bid, October corn?
-> **ELEVATOR:** Forty-five under the December.
-> **FARMER:** That's ugly.
-> **ELEVATOR:** I've got eleven days of receiving and nine months of crop coming at me. Forty-five under is me telling you not to bring it.
-> **FARMER:** And if I store it with you instead?
-> **ELEVATOR:** Then I charge you rent instead of paying you a bid. Same conversation, other direction.
-
-A bid can be a refusal. Forty-five under is not an opinion about corn; it is a statement about how much room is left in the house. When space is the binding constraint, the posted bid stops functioning as a price and starts functioning as a queue-management tool — and a merchant who reads country bids as market sentiment will get the direction of the next basis move exactly wrong.
-
-The second half of the exchange matters as much as the first. The elevator offers the farmer a choice between a bad bid and a storage tariff. Those are the same transaction viewed from opposite ends: in one, the elevator buys the grain cheaply and owns the space problem; in the other, the farmer keeps the grain and rents the space problem. The elevator is indifferent to which, because it is selling the same thing either way.
-
-### The trade, worked
-
-Take the elevator's side of it, using Friday's December corn at $5.08½.
-
-| Step | | |
-|---|---|---|
-| Harvest: buy cash corn | 45 under Dec | $4.63½ |
-| Sell December futures | | flat price hedged |
-| Late February: sell cash | 15 under Mar | |
-| Assume Mar–Dec spread | +18¢ | |
-
-The trick that makes this readable is to restate both ends against a single month. Fifteen under March, when March is eighteen over December, is **three cents over December**.
-
-| | vs December |
-|---|---|
-| Basis bought | −45¢ |
-| Basis sold | +3¢ |
-| **Gross gain** | **48¢/bu** |
-| Interest, $4.63½ at 5% for 4 months | −7½¢ |
-| Shrink, drying, turning the pile | −3¢ |
-| **Net** | **37½¢/bu** |
-
-On a three-million-bushel house that turns its space once, that is a little over **$1.1 m**. The elevator never had a view on corn. Not once. It was paid for owning room in the six weeks of the year when nobody else had any.
-
-Which gives the sentence worth keeping: **the carry in the futures curve can only be collected by somebody who has a bin.** Ep 3 said a carry market pays you to store. That was half of it. It pays you to store *if you have somewhere to store*. If you do not, the carry is a number on a screen that somebody else is going to earn, and the store-or-sell decision that looked like an optimisation was never available to you at all.
-
-### The shed that was never built
-
-Now put the fifteen billion bushels back on the table.
-
-| | Bushels |
-|---|---|
-| Total US grain storage capacity, 2025 | ~25.3 bn |
-| Same, 2019 | ~25.0 bn |
-| On the 2000–2019 trend | 27.5 bn |
-| On-farm | 13.6 bn |
-| Off-farm commercial | 11.9 bn |
-
-Between 2000 and 2019 the US added an average of 349 m bu of capacity a year. Since 2020 it has added essentially nothing — about 337 m bu in six years, which is less than one year of the old trend. Production did not stop growing.
-
-```chart
-{"type":"bar","unit":"billion bushels","title":"The shed that was never built",
- "x":["2019 actual","2025 actual","2025 on trend"],
- "series":[{"name":"US grain storage capacity","values":[25.0,25.3,27.5]}],
- "caption":"Six years of building added 0.3 bn bushels where the previous two decades' trend would have added 2.5. That missing capacity is not an abstraction — it is why harvest basis goes where it goes, and who gets paid when it does.",
- "source":"farmdoc daily, University of Illinois, February 2026, from USDA capacity series."}
-```
-
-The consequences show up in the utilisation numbers rather than the capacity numbers. On 1 December 2025 on-farm storage was running at **80% of capacity**, a record, and off-farm at 65%. System-wide surplus capacity was about **5%**, against a century average of 15%. That is the tightest since 1988.
-
-Five percent is not a cushion. It is a rounding error, and it means the marginal bushel in a good year has no home at any basis — which is precisely when temporary storage appears: ground piles, bunkers and bags, cheap per bushel to build and expensive per bushel in spoilage, rehandling and quality loss. The cost of the pile is what sets the floor under how wide the basis has to go.
-
-### Form: the cheapest transformation there is
-
-Ep 2 said merchants are paid for three transformations: space, time and form. Freight is space. Storage is time. Blending is form — and it is the cheapest form change in the business. No chemistry, no plant, no conversion cost worth the name. Just a weighted average and two tanks.
-
-Take a 60,000 t cargo of milling wheat sold against a 12.0% protein minimum.
-
-| Lot | Tonnes | Protein | Cost $/t |
-|---|---|---|---|
-| A | 40,000 | 12.4% | 250.00 |
-| B | 20,000 | 11.2% | 232.00 |
-| **Blend** | **60,000** | **12.0%** | **244.00** |
-
-Lot B is off-spec and nobody wants it, which is why it was $18 cheap. Blended, the cargo is exactly on spec and worth the full $250.00/t.
-
-```chart
-{"type":"waterfall","unit":"USD per tonne","title":"Blending a cargo onto spec",
- "steps":[{"label":"Sale value, 12.0% cargo","value":250,"kind":"base"},
-          {"label":"Weighted cost of blend","value":-244},
-          {"label":"Elevation and handling","value":-3},
-          {"label":"Margin","kind":"total"}],
- "caption":"Neither lot could be sold as a 12% cargo alone. Six dollars a tonne appears out of a weighted average, and three of it survives the handling cost — $180,000 on the cargo, created by arithmetic.",
- "source":"Worked example, episode 11."}
-```
-
-Six dollars a tonne out of nothing but a weighted average. Take $3.00/t for elevation and handling and $3.00/t survives: **$180,000** on the cargo.
-
-The discount lot was cheap for a specific reason worth naming. Most buyers physically cannot use 11.2% wheat — their contracts, their mills or their customers will not take it. That thins the bidding to the handful of operators with segregated space and a blending sale to put it into, which is why the blender frequently *sets* the discount rather than taking it. Blending value is not really a quality trade. It is a trade on being one of very few people able to bid.
-
-### The part that gets people fired
-
-Protein averages. Moisture averages. Test weight averages.
-
-Aflatoxin does not average. Neither does a live insect infestation, nor an unapproved genetic event, nor — as a practical matter — a falling number.
-
-These are not quality specifications with a discount schedule attached. They are **pass/fail conditions on the whole lot**. Pour 10,000 t of contaminated corn into a 50,000 t bin of clean corn and you have not diluted anything. You have 60,000 t of contaminated corn, and you have converted a containable problem into a cargo-sized one.
-
-Falling number deserves its own line because it looks like a number that should average and does not. The Hagberg test measures how fast a flour-and-water slurry loses viscosity, which is a proxy for alpha-amylase activity in sprout-damaged grain. Enzymes are catalysts, so a small weight share of highly active material degrades starch far out of proportion to its tonnage, and the blend lands below the weighted mean — always in the direction that costs the seller. Ep 5 priced the consequence: a low falling number demotes milling wheat to feed at roughly $40/t.
-
-And ep 4's rule closes the loop: the quality certificate is final at load. The blend has to be right before the grain is on the ship, because after it is, the arithmetic is no longer a commercial question. It is a claim.
-
-### Flows, replacement value, and the bottleneck
-
-The last piece ties the two halves together.
-
-Traders talk about **replacement value**: what it would cost, right now, to buy back what has just been sold. It is a better question than whether the price was good, because it is the only one that answers whether the trade can be done again tomorrow. A sale at a record price into a market where the replacement is unavailable at any price is not a good trade. It is the end of a business line.
-
-Replacement value is set by whoever controls the bottleneck. In a congested port during harvest week, the operator with space names the price and everybody else takes it — and no view on flat price competes with that. This is also why elevation margin is quietly counter-cyclical to trading margin: the weeks when the basis is at its worst and the trading book is grinding are exactly the weeks when the space is worth the most.
-
-Ep 10 closed by asking why merchants rent ships but own elevators. That is the answer. A ship is one of many; if the owner will not fix at your number, there is another vessel behind it, and the Baltic assessment tells you roughly what it should cost. A berth in a congested terminal in the week everybody needs it is one of one, and there is no index for it because there is no substitute to average against.
-
-Which is the same shape as the Russian story in this morning's pulse, at national scale. Something like 140 mt of grain exists, the world wants it, and more than 90% of the export capacity that would move it is offline. The crop did not shrink. The bottleneck closed. And the price of the same wheat split into two prices — a world price going up, and a domestic price at roughly 12,000 roubles a tonne going down — with storage and the ability to move sitting in the gap between them.
-
-That gap is the business.

package/ep11.script.txt

@@ -1,107 +0,0 @@
-Last year the United States built almost no new grain storage. ||| 0.4
-This autumn it is going to harvest something like fifteen billion bushels of corn. ||| 0.5
-Those two facts meet in September. And the place where they meet has a name. It is called the basis. ||| 0.7
-This is Soft Commodity Trading, episode eleven. Storage, elevation, and how grain actually flows. ||| 0.8
-Friday first. December corn settled at five dollars eight and a half, up five cents, a two and a half year high. ||| 0.4
-On the week it added twenty five and a quarter cents. ||| 0.5
-November beans finished at twelve dollars thirty nine and a half, up three, and up forty seven cents on the week. ||| 0.5
-Wheat did nothing. Chicago September down a cent and a quarter at six eighty one and a half. Kansas City down six at seven fifty six and a quarter. ||| 0.5
-Meal up two dollars. Oil down a hundred and eighty three points. ||| 0.6
-The number that moved corn came out after the bell on Friday. Pro Farmer finished its crop tour. ||| 0.4
-It put the national corn yield at a hundred and seventy three point two bushels an acre. ||| 0.4
-U S D A said a hundred and eighty point seven three weeks ago. ||| 0.5
-That is seven and a half bushels of daylight between the two, and it is why December corn is at a two and a half year high. ||| 0.6
-The beans went the other way. Pro Farmer has fifty three point three against U S D A's fifty two point seven. Slightly bigger, not smaller. ||| 0.7
-Now the part that matters for today. Even at Pro Farmer's number, the corn crop is over fifteen billion bushels. ||| 0.4
-That crop is arriving into a storage system that stopped growing six years ago. Hold that thought. ||| 0.7
-The geopolitical read, and it is the same subject. ||| 0.4
-More than ninety percent of Russia's grain export capacity in the Azov and Black Sea basin is currently offline. ||| 0.5
-Three terminals at Novorossiysk suspended last week. Taman stopped in late July. Navigation in the Sea of Azov has been suspended since July. ||| 0.5
-One deepwater grain terminal is still working in the region. That basin moved forty six million tonnes last season. ||| 0.6
-Here is the mechanism, and it is not the one you would reach for first. ||| 0.4
-When an export door shuts, the grain does not disappear. It queues. ||| 0.5
-Russia has just harvested about a hundred and forty million tonnes. Exporters have stopped buying it, because they cannot ship it. ||| 0.5
-So it backs up inland. The elevators fill. And the domestic price falls off a cliff. ||| 0.5
-Fourth class Russian wheat is around twelve thousand roubles a tonne. A year ago it was fifteen thousand. ||| 0.6
-That is the whole lesson in one sentence. The world price of wheat is going up, and the Russian farmer's price is going down, at the same time, in the same crop. ||| 0.6
-What sits between those two prices is storage and the ability to move. ||| 0.8
-So. Two things today. What an elevator actually sells. And what blending is really worth. ||| 0.7
-Start with the elevator, because almost everybody gets it wrong. ||| 0.4
-An elevator is not a warehouse. It is not a speculator either. ||| 0.5
-It buys grain at the posted bid, and it sells the board against it within minutes. Flat price gone. ||| 0.5
-What it owns is space, and the right to move grain through that space. ||| 0.6
-And here is the unit moment. Commercial storage is quoted in cents per bushel per month, sometimes per day. ||| 0.4
-Shrink is quoted in percent, and it is the weight that disappears when wet grain is dried down to a keeping moisture. ||| 0.4
-Both of those are costs the elevator either charges out or eats. Neither of them is a price view. ||| 0.6
-Listen to how a bid actually gets given in October. ||| 0.5
-FARMER: What are you bid, October corn? ||| 0.25
-ELEVATOR: Forty five under the December. ||| 0.25
-FARMER: That's ugly. ||| 0.25
-ELEVATOR: I've got eleven days of receiving and nine months of crop coming at me. Forty five under is me telling you not to bring it. ||| 0.3
-FARMER: And if I store it with you instead? ||| 0.25
-ELEVATOR: Then I charge you rent instead of paying you a bid. Same conversation, other direction. ||| 0.7
-Notice what happened. A bid can be a refusal. ||| 0.5
-Forty five under is not an opinion about corn. It is a statement about how much room is left. ||| 0.5
-When space is the binding constraint, the bid stops being a price and starts being a queue management tool. ||| 0.7
-Now the trade the elevator is actually doing. Work it through. ||| 0.5
-Harvest. It buys corn at forty five under the December. With December at five oh eight and a half, that is four dollars sixty three and a half cash. ||| 0.4
-It sells December futures against it. Flat price is dead. ||| 0.5
-Four months later it sells that corn at fifteen under the March. ||| 0.4
-And say March is trading eighteen cents over December. ||| 0.5
-Fifteen under March, when March is eighteen over December, is three cents over December. ||| 0.6
-So measure both ends against the same month, and the whole thing gets simple. ||| 0.4
-It bought the basis at forty five under. It sold the basis at three over. ||| 0.4
-Forty eight cents a bushel. ||| 0.6
-Against that, the costs. Interest on four dollars sixty three for four months, at five percent, is about seven and a half cents. ||| 0.4
-Shrink, drying and turning the pile, call it another three. ||| 0.4
-Eleven cents of cost. Thirty seven cents a bushel left. ||| 0.6
-On a three million bushel country elevator that turns its space once, that is a little over a million dollars. ||| 0.6
-And now the point. The elevator never had a view on corn. Not once. ||| 0.5
-It got paid for owning room in the six weeks of the year when nobody else had any. ||| 0.6
-Which gives you the sentence worth keeping from all of this. ||| 0.4
-The carry in the futures curve can only be collected by somebody who has a bin. ||| 0.6
-We said in episode three that a carry market pays you to store. That was half of it. ||| 0.4
-It pays you to store if you have somewhere to store. If you do not, the carry is just a number on a screen that somebody else is going to earn. ||| 0.8
-Which brings us back to the fifteen billion bushels. ||| 0.5
-Total American grain storage capacity is about twenty five point three billion bushels. ||| 0.4
-In two thousand nineteen it was just over twenty five. Six years, three hundred million bushels added. ||| 0.5
-On the previous twenty year trend it would have been twenty seven and a half billion by now. ||| 0.5
-That missing two point two billion bushels of shed is the reason harvest basis behaves the way it does. ||| 0.6
-On the first of December last year, on farm storage was running at eighty percent of capacity, which is a record. ||| 0.4
-Surplus capacity across the whole system was about five percent. The average over this century is fifteen. ||| 0.5
-Five percent is the tightest it has been since nineteen eighty eight. ||| 0.7
-Second thing. Form. ||| 0.5
-Episode two said merchants get paid for three transformations. Space, time, and form. ||| 0.4
-Storage is time. Freight is space. Blending is form, and it is the cheapest form change in the entire business. ||| 0.5
-No chemistry. No plant. Just arithmetic and two tanks. ||| 0.6
-Here is the example. A cargo of sixty thousand tonnes of milling wheat, sold against a twelve percent protein minimum. ||| 0.5
-The blender has forty thousand tonnes of twelve point four protein, which cost him two hundred and fifty dollars a tonne. ||| 0.4
-And he has twenty thousand tonnes of eleven point two protein. Off spec. Nobody wants it. He paid two hundred and thirty two. ||| 0.6
-Blend them. Forty thousand at twelve point four, twenty thousand at eleven point two. ||| 0.4
-The weighted average is exactly twelve point zero. ||| 0.6
-The cargo is now on spec. It is worth two hundred and fifty dollars a tonne. ||| 0.4
-His weighted cost was two hundred and forty four. ||| 0.5
-Six dollars a tonne, out of nothing but a weighted average. Take three dollars off for elevation and handling and he keeps three. ||| 0.4
-On sixty thousand tonnes, a hundred and eighty thousand dollars. ||| 0.7
-And the discount lot was cheap for a specific reason. Most buyers physically cannot use it. ||| 0.5
-Which means the blender is frequently the only real bidder for it, and that is why he gets to set the discount rather than take it. ||| 0.6
-Now the part that gets people fired. ||| 0.5
-Protein averages. Moisture averages. Test weight averages. ||| 0.4
-Aflatoxin does not average. Neither does a live insect infestation, and neither does an unapproved genetic event. ||| 0.6
-Those are not quality specifications. They are pass or fail conditions on the whole lot. ||| 0.5
-Pour ten thousand tonnes of contaminated corn into a fifty thousand tonne bin and you do not have a dilution. You have sixty thousand tonnes of contaminated corn. ||| 0.6
-Remember from episode four that the quality certificate is final at load. ||| 0.4
-So the blend has to be right before the grain is on the ship, not after. ||| 0.8
-Last piece, quickly, because it ties the two halves together. ||| 0.4
-Traders talk about replacement value. It means, what would it cost me right now to buy back what I just sold. ||| 0.5
-It is a better question than whether the price was good, because it is the only one that tells you whether you can do it again tomorrow. ||| 0.6
-And replacement value is set by whoever controls the bottleneck. ||| 0.5
-In a congested port, in a harvest week, the person who owns the space names the price, and no view on the flat price competes with that. ||| 0.6
-Episode ten ended by asking why merchants rent ships but buy elevators. ||| 0.4
-This is the answer. A ship is one of many. A berth in a congested terminal, in the week everybody needs it, is one of one. ||| 0.8
-So, three things. ||| 0.4
-The elevator is not long grain. It is long space, and the basis is what space costs. ||| 0.5
-The carry only belongs to somebody with a bin, which is why a storage shortage shows up as a price, not as a queue. ||| 0.5
-And blending creates real money out of arithmetic, right up to the point where the specification is pass or fail, at which point it destroys real money just as fast. ||| 0.7
-Tomorrow, we move into softs. Coffee. Arabica against robusta, and why the same word covers two completely different markets. ||| 0.5
-The quiz is in the notes, with the answers underneath. Take it before you look. ||| 0.7