@sdelsad/commodity-desk-daily 1.0.26 → 1.0.28

package/ep08.md

@@ -1,301 +0,0 @@
-## Market pulse
-
-**Beans went nowhere and soybean oil fell out of bed — so the crusher's margin moved while the seed did not.**
-
-| Commodity | Contract | Price | Change |
-|---|---|---|---|
-| Corn | Sep (CBOT) | 463¼ c/bu | −1¾¢ |
-| Corn | Dec (CBOT) | 488 c/bu | −1½¢ |
-| Soybeans | Sep (CBOT) | 1200¾ c/bu | −¼¢ |
-| Soybeans | Nov (CBOT) | 1216¾ c/bu | +¾¢ |
-| Soymeal | Sep (CBOT) | — | +0.25% |
-| Soyoil | Sep (CBOT) | — | −2.5% |
-| Wheat SRW | Sep (CBOT) | 664½ c/bu | −10¼¢ |
-| Wheat HRW | Sep (KC) | 743¾ c/bu | −15¢ |
-
-Tuesday was a profit-taking session after a strong week, and the damage landed on wheat. Chicago soft red gave back 10¼ cents and Kansas City hard red gave back fifteen, close to 2%. Corn drifted lower on ratings that fell a point to 60% good to excellent, with 76% of the crop at dough — ahead of the five-year average of 70%. Soybeans held: ratings slipped to 61%, 85% of the crop is setting pods, and private exporters reported another 5.0 m bu sold to China for 2026/27.
-
-The move worth reading was inside the bean complex itself. Meal added a quarter of a percent. Oil lost nearly 2.5%. Beans finished flat. A soybean is not one price, and on Tuesday two of its three prices went in opposite directions.
-
-**The geopolitical read: policy is now half of a soybean.** Soybean oil no longer prices as a food. It prices off American biofuel rules — the mandated renewable fuel volumes and the clean fuel production credit the trade calls 45Z — and off Brent, which sat just under $91. When those rules are uncertain, oil trades like a fuel, and it drags the crush with it. The second lever is fiscal. Argentina is cutting export taxes on a published schedule: soybeans at 24%, falling to 21% by end-2027 and 15% by end-2028, with meal and oil taxed *below* the bean. That gap is deliberate. It taxes the export of a seed more heavily than the export of a product, which is a subsidy for crushing at home. Two governments, two instruments, one margin.
-
-```chart
-{"type":"bar","unit":"% change, Tue 18 Aug",
- "title":"One complex, two directions",
- "x":["Corn Dec","Beans Nov","Meal Sep","Oil Sep","Wheat Sep","KC Sep"],
- "series":[{"name":"Change on the day","values":[-0.31,0.06,0.25,-2.50,-1.52,-1.98]}],
- "caption":"Soybeans finished unchanged while soybean oil lost 2.5%. A flat bean price is not a flat day for anyone who owns a crush plant. Wheat took the profit-taking.",
- "source":"CBOT and KC settlements, Tuesday 18 August 2026, from the daily market recap."}
-```
-
-## Key takeaways
-
-- One bushel of soybeans is 44 lb of meal and 11 lb of oil. The crusher does not choose the ratio, so he is a price-taker on the mix and cannot overweight the market he likes.
-- The two multipliers are the whole of the arithmetic: meal price × 0.022, oil price × 0.11, minus the bean price. Everything else is detail.
-- Oil is now more than half the gross product value. Meal used to be two-thirds of it. A crusher's biggest single exposure is to fuel policy, not to agriculture.
-- The board crush is a quote assembled from three futures prices. The plant crush is that number plus three separate basis positions, minus 35–50¢/bu of conversion cost.
-- Crushing capacity takes two to three years to build, so a wide margin is not competed away by new plants. It is competed away through the bean basis at the gate, which is why the screen can show a fat crush that nobody is earning.
-- A negative board crush rarely stops a plant, because the meal is already sold and restarting costs days. The real option is on variable margin over cash costs, on uncommitted volume only.
-- The margin is small relative to what it is built from. A 1% move in the products is about 6% of the crush.
-
-## Vocabulary
-
-| Term | Meaning |
-|---|---|
-| **Board crush** | The processing margin implied purely by futures prices, meal × 0.022 plus oil × 0.11 minus the bean price, in dollars per bushel |
-| **Plant crush** | What a physical plant actually earns, the board crush adjusted for bean, meal and oil basis and net of conversion cost |
-| **Gross processing margin (GPM)** | The industry name for product value minus raw material cost, the crush stated as a margin |
-| **Putting on the crush** | Buying bean futures and selling meal and oil futures against them, in a 10-11-9 lot ratio, which fixes the margin |
-| **Reverse crush** | The opposite position, short beans and long products, put on when a processor expects to idle capacity rather than run it |
-| **Oil share** | Soybean oil's percentage of the combined value of meal and oil out of one bushel |
-| **Meal contract** | CBOT soybean meal, 100 short tons, quoted in dollars per short ton |
-| **Oil contract** | CBOT soybean oil, 60,000 lb, quoted in cents per pound |
-| **Conversion cost** | The variable cost of turning beans into products — gas, power, hexane, labour, maintenance — typically 35–50 c/bu at a modern plant |
-| **Crush capacity** | Installed daily processing volume, a physical constraint that cannot be expanded inside a marketing year |
-| **Run rate** | The share of installed capacity a plant is actually operating at, the lever a crusher pulls when margins move |
-| **Hexane** | The solvent used to extract the last of the oil from the flaked bean, and a real line in the conversion cost |
-| **Joint product** | Two outputs produced in fixed proportion from one input, so that neither can be made without the other |
-| **45Z** | The US clean fuel production credit, one of the two policy levers that sets American soybean oil demand |
-| **Draw area** | The geographic catchment a plant buys its beans from, whose size sets how hard it must bid the local basis |
-
-## Quiz
-
-**Q1.** Compute one from scratch. March meal is $325.00/short ton, March oil is 70.10 c/lb and March beans are 1248 c/bu. Give the board crush in dollars per bushel, in dollars per tonne of beans, and give the oil share. Then say which of the three legs you would hedge first if you could only reach one of them before the close, and why.
-
-**Q2.** A plant runs 165,000 bu/day. The board crush is $2.52. Its bean basis at the gate is +18¢ over November, its meal basis is $6.00/short ton *under* December, its oil basis is 0.40 c/lb over December, and variable conversion cost is 42¢/bu. Compute the plant's actual margin per bushel and per day. Then explain which of those four adjustments is the one that moves most from week to week, and what that implies about where a crush trader should spend their attention.
-
-**Q3.** The board crush goes to −15¢/bu and stays there for three weeks. Your plant has already sold 70% of next month's meal production forward at fixed prices, and shutting the line down costs roughly $400,000 plus four days. Argue the case for running anyway, then argue the case for cutting the run rate to 60%. State the single number that decides it, and explain why a trader watching only the board crush would reach the wrong answer.
-
-**Q4.** *(Ep 7)* A desk rebuilds the US soybean sheet and lands on a carryout 12% below USDA's, having used a yield only 1.1% below USDA's. Ep 7 established that a carryout gap is usually half a crop view and half a demand view. Decompose this one: what has the desk almost certainly done to the demand side, and which single line is the most likely home for it? Then say why crush demand in particular makes this an unusually loaded disagreement in soybeans compared with corn.
-
-**Q5.** *(Ep 7)* Feed and residual absorbs measurement error as well as livestock demand. Soybeans do not have a feed and residual line of the same character, because the crop is dominated by one measurable use. Name that use, explain why it makes the soybean balance sheet tighter to argue about than corn's, and identify where the residual uncertainty in soybeans actually hides instead.
-
-**Q6.** *(Ep 5)* You are short 1,200 Matif lots against a Black Sea wheat cargo — the cross-hedge from ep 5. Your risk manager now asks you to hedge a 60,000 t Brazilian soybean cargo the same way, using CBOT beans. Explain why the second hedge is a fundamentally better one than the first, in terms of what each contract is actually referencing, and name the residual exposure the bean hedge still leaves you with.
-
-**Q7.** *(Ep 5)* Ep 5 valued the KC-over-Chicago spread of 68 c/bu as $26/t of protein. On Tuesday KC September closed at 743¾ and Chicago September at 664½. Compute the spread in cents and in dollars per tonne. It has widened since ep 5 — give two distinct explanations, one about protein and one about export competitiveness, and say what you would look at to tell them apart.
-
-**Q8 — Conversion drill.** A crush plant's draw area covers 640,000 hectares of soybeans. Convert that to acres. At an average 3.4 t/ha, convert the production to bushels. The plant runs 165,000 bu/day and operates 330 days a year. How many years of throughput does its draw area produce, and what does that ratio tell you about how hard it will have to bid the local basis?
-
-## SOLUTIONS (spoilers)
-
-**A1.** The arithmetic, in the fixed order.
-
-| Leg | Price | Multiplier | Value per bushel |
-|---|---|---|---|
-| Meal | $325.00/short ton | × 0.022 | $7.150 |
-| Oil | 70.10 c/lb | × 0.11 | $7.711 |
-| **Gross product value** | | | **$14.861** |
-| Beans | 1248 c/bu | | −$12.480 |
-| **Board crush** | | | **$2.381** |
-
-So **$2.38/bu**. Per tonne, multiply by the 36.744 bushels in a tonne of soybeans: **$87.50/t**. Oil share is 7.711 ÷ 14.861 = **51.9%**.
-
-Which leg to hedge first: **the beans**. It is the largest single number in the calculation by a wide margin — $12.48 against $7.15 and $7.71 — so an unhedged bean leg carries more variance than either product leg on its own. The instinct to reach for oil first, because oil is the volatile one, is the trap. Volatility matters, but it is volatility *times notional*, and the bean leg's notional is 1.6× either product's. There is a second, practical reason: bean futures are the most liquid of the three, so it is the leg you can actually get done in size in the last minutes of a session. Hedge the thing you can hedge.
-
-**A2.** Build it as a bridge from the board.
-
-| Line | ¢/bu |
-|---|---|
-| Board crush | +252.0 |
-| Bean basis paid at the gate (+18¢) | −18.0 |
-| Meal basis ($6.00/st under × 0.022) | −13.2 |
-| Oil basis (0.40 c/lb over × 0.11) | +4.4 |
-| Conversion cost | −42.0 |
-| **Plant crush** | **+183.2** |
-
-Note the sign convention, which is where most people get this wrong. A bean basis *over* futures is a cost, because the plant is buying. A meal basis *under* futures is also a cost, because the plant is selling. Both work against you here; only the oil basis helps.
-
-$1.832/bu × 165,000 bu/day = **$302,280 a day**, about $6.3 m a month on a 21-day month.
-
-The line that moves most week to week is the **bean basis**. Conversion cost is close to fixed over a quarter. Meal and oil basis move, but within relatively narrow ranges set by freight to the feed mill and the refinery. The bean basis is the competitive variable: it is where the plant fights other plants, the export elevator and the farmer's willingness to sell, and it can move 20–30¢ in a fortnight when the board crush is wide and everyone is bidding for the same beans.
-
-The implication is uncomfortable for anyone who came from a screen-trading background. The board crush is the number on everyone's monitor, and it is the number a crush trader has the *least* ability to influence. The bean basis is the number that decides whether the plant makes money, and it is the one the desk actually controls, one origination decision at a time. Attention should be roughly inverse to how visible the number is.
-
-**A3.** **The case for running.** 70% of next month's meal is sold forward at fixed prices. That meal has to come from somewhere. If the line stops, the plant must buy meal in the market to honour those sales, at whatever price prevails — and in a negative-crush environment, meal is exactly the product that rallies as run rates fall industry-wide. The plant would be buying back its own shortfall into a market its own shutdown helped tighten. Add the $400,000 and four days, and add the fact that a stopped line means the bean book, storage and staff do not stop costing money.
-
-**The case for cutting to 60%.** A negative margin multiplied by volume is a loss that scales linearly. If the committed meal is 70% of a full run, then running at 60% still covers most of it while crushing 40% fewer bushels at a loss. Three weeks at −15¢ on 165,000 bu/day is roughly $520,000 of board-level loss at full rate; cutting to 60% saves about $210,000 of it, which is comparable to the shutdown cost without incurring it.
-
-**The single number that decides it: the variable margin over cash costs on the uncommitted volume.** Not the board crush. The committed 70% is not a decision any more — that meal is sold, and the only question is whether it is cheaper to make it or buy it. The decision lives entirely in the remaining 30%. If, on that marginal volume, revenue at today's cash meal and oil prices exceeds the cash cost of the beans plus variable conversion, the plant runs it. If not, it cuts. Fixed costs and the board crush are both irrelevant to that comparison.
-
-The trap the question sets is that a trader watching only the board crush sees −15¢, concludes the industry is losing money, and shorts meal or buys beans. The actual chain runs the other way: negative margins cut run rates, cut run rates tighten meal supply, tight meal supply rallies meal, and the crush repairs itself. A deeply negative crush is more often a reason to own meal than to sell it.
-
-**A4.** A 1.1% yield cut with a 12% carryout cut is far more leverage than the supply side alone can produce. In ep 7's US corn arithmetic, roughly a 1% yield error produced something on the order of a 10% carryout error, and even that ratio needs a large crop sitting on a small carryout. Here the desk has done something to demand as well: it is **carrying more use than USDA**, and the most likely home for it is the **crush** line, with exports the second candidate.
-
-Crush makes this loaded in a way corn's disagreements are not. In corn, the two big demand lines — feed and ethanol — are either inferred or set by policy and capacity that changes slowly. In soybeans, crush is a *margin-driven* line. If the board crush is wide, plants run harder, and crush demand rises endogenously. So a desk that is bullish the crush margin is, by construction, bullish crush volume, which cuts the carryout, which is itself bullish beans, which compresses the crush margin. The demand line and the price feed back into each other inside the same sheet. That is why two competent soybean analysts can agree on the crop to within a bushel and still be 100 m bu apart on the carryout.
-
-**A5.** The use is **crush**, and in the US it is the largest single domestic use of the soybean crop by a wide margin. It is *measured*, not inferred: NOPA publishes a monthly crush figure from its member plants, and members account for the large majority of US capacity. That is a monthly, hard, published number against a line that in corn would be a quarterly inference.
-
-This makes the soybean sheet tighter to argue about. There is less room to hide a mistake, because a wrong crush assumption is contradicted by a real print within weeks, rather than surviving until the next Grain Stocks survey.
-
-The residual uncertainty does not vanish, though — it **moves to exports**, and specifically to the gap between commitments and shipments. Ep 7's pulse made the point about the buyer's clock: a Chinese purchase is a promise on a balance sheet, and a loading is a fact on a vessel. Sales can be booked, rolled, switched to another origin or cancelled. So the soybean sheet's soft line is not a residual absorbing measurement error, it is a demand line absorbing *counterparty behaviour*. Different problem, same effect on the carryout.
-
-**A6.** The ep 5 Matif hedge was a **cross-hedge**: 12.5% Russian milling wheat, FOB Black Sea, hedged with a contract that references EU milling wheat delivered Rouen–Dunkirk. Different wheat, different quality spec, different delivery geography, different currency. The contract is a proxy for a world price, and the basis between the two is itself a large, volatile, unhedgeable position — which is how the worked example produced $7.50/t of slippage on a 10 EUR/t European move.
-
-CBOT beans against a Brazilian cargo is a materially better hedge for one structural reason: **soybeans are close to a globally fungible commodity, and CBOT is the world's reference price for it.** Brazilian beans and US beans are substitutes into the same crushers, with a protein and oil-content difference that is small and slow-moving. Brazilian physical trades explicitly as a differential *to CBOT* — Paranaguá plus or minus so many cents against a named month — which is the clearest possible evidence that the contract is the right reference. The Black Sea has no futures at all, only assessments, which is why ep 5 had to reach for Matif in the first place.
-
-The residual exposure is the **basis itself**: the Paranaguá or Santos differential against the board. Freight, the Brazilian farmer's selling pace, the real, line-ups at the port and Chinese demand all move it, and none of them are in the CBOT price. That is not a flaw in the hedge — it is the trade. The hedge is meant to remove flat price and leave the basis, and per ep 2, the basis is what the merchant is paid to be long or short of. The difference from the wheat case is one of scale: a bean basis position is a known, tradeable exposure of a few tens of cents; the Black Sea-to-Matif basis was an unknown of $7.50/t on a quiet move.
-
-**A7.** 743.75 − 664.50 = **79¼ c/bu**.
-
-Convert: wheat is 36.744 bu/t, so 0.7925 × 36.744 = **$29.12/t**. Against ep 5's 68¢ and $26/t, the spread has widened about 11¼¢, or roughly $3/t.
-
-**Explanation one, protein.** KC prices hard red winter at 11–12.5% protein; Chicago prices soft red at around 10%. The spread is the market's price for those extra protein points. It widens when the milling market is short of protein — a low-protein HRW harvest, or strong flour demand for bread grists rather than biscuit grists.
-
-**Explanation two, export competitiveness.** HRW is the US export wheat; SRW is more of a domestic and Gulf-of-Mexico feed-and-biscuit wheat. If US HRW is winning tenders — or if a competing origin's HRW-substitute is unavailable — KC gets bid on export demand alone, with nothing to do with protein.
-
-**How to tell them apart.** Look at the cash protein scale first: the premium paid for 12% over 11% HRW at Gulf and at the plains elevators. If protein is the story, that ladder steepens and the KC spread widens with it. If the ladder is flat and KC is still bid, it is export demand, and you would confirm it in the weekly export sales, in the tender results, and in the KC cash basis at the Gulf. On Tuesday the direction argues for neither, incidentally: KC fell *harder* than Chicago, which narrows the spread on the day, so this was profit-taking on a position rather than a change in the underlying story.
-
-**A8 — Conversion drill.** 640,000 ha → acres. Fast method: ×2.5 and shave 1%. 640,000 × 2.5 = 1,600,000, less 16,000 = **1,584,000 acres**, call it 1.58 m acres. (Exact: 640,000 × 2.47 = 1,580,800.)
-
-Production: 640,000 ha × 3.4 t/ha = **2,176,000 t**. In bushels, × 36.744 = **79.96 m bu**, call it 80 m.
-
-Plant throughput: 165,000 bu/day × 330 days = **54.45 m bu a year**.
-
-80.0 ÷ 54.45 = **1.47 years** of throughput sitting in the draw area.
-
-What it tells you: the plant needs roughly **68% of every soybean grown in its catchment**. That is a demanding share, and it is why the bean basis at the gate is the number that decides the plant's margin. It cannot simply wait for beans to arrive; it has to outbid the export elevator, the river terminal and the farmer's storage decision for two bushels in every three, all year, every year. A plant with a 3× ratio can be relaxed about basis. A plant at 1.5× cannot, and that is the physical reason a wide board crush gets competed away through the origination bid rather than through new capacity.
-
-## The episode, in writing
-
-### One seed, three markets
-
-A soybean is not a commodity. It is a package of two commodities that have to be sold together.
-
-Put 60 lb of beans — one bushel — through a crush plant and you get roughly **44 lb of meal** and **11 lb of oil**, with the balance lost to hulls and moisture. Those proportions are set by the seed, not by the plant, and they do not respond to which market a trader would rather be in.
-
-That matters because meal and oil are not related businesses.
-
-**Meal is a protein market.** It goes into a feed ration, where it competes with fishmeal, rapeseed meal and — at the margin, on an energy-versus-protein trade-off — with corn. Its demand is livestock: hog herds in China, poultry in Brazil and Southeast Asia, dairy in Europe. It is agricultural, seasonal and slow.
-
-**Oil is a vegetable oil market with an energy problem attached.** It competes with palm and canola in a fryer and with diesel in a tank. Its demand is set as much by biofuel mandates and tax credits as by cooking. It is fast, political, and correlated to crude.
-
-The crusher is exposed to both, in a fixed ratio, permanently. He is a price-taker on the mix. That single fact generates everything else in this lesson.
-
-### The two multipliers
-
-Three markets means three quoting conventions, and the arithmetic does not work until they are collapsed into one unit.
-
-| Instrument | Contract size | Quoted in |
-|---|---|---|
-| Soybeans (ZS) | 5,000 bu | cents per bushel |
-| Soybean meal (ZM) | 100 short tons | dollars per short ton |
-| Soybean oil (ZL) | 60,000 lb | cents per pound |
-
-The collapse is two constants:
-
-- **Meal:** 44 lb out of a 2,000 lb short ton is 0.022. Meal price × 0.022 = meal value per bushel.
-- **Oil:** 11 lb, priced in cents, divided by 100, is 0.11. Oil price × 0.11 = oil value per bushel.
-
-Add the two, subtract the bean price, and you have the **board crush**. Those two numbers, 0.022 and 0.11, are worth committing to memory. Every conversation about this market runs through them.
-
-### The board crush, on Tuesday's numbers
-
-| Leg | Price | Multiplier | $/bu |
-|---|---|---|---|
-| Meal, Dec | $321.60/short ton | × 0.022 | 7.08 |
-| Oil, Dec | ≈69.2 c/lb | × 0.11 | 7.61 |
-| **Gross product value** | | | **14.69** |
-| Beans, Nov | 1216¾ c/bu | | −12.17 |
-| **Board crush** | | | **2.52** |
-
-The oil figure is derived rather than quoted: December oil closed Monday at 70.96 c/lb, and Tuesday's session took the front month down about 2.5%, which puts December near 69.2.
-
-**$2.52 a bushel.** In the units used outside the United States, multiply by the 36.744 bushels in a tonne of soybeans: **$92.50 per tonne of beans crushed**.
-
-```chart
-{"type":"waterfall","unit":"$/bu","title":"The board crush, decomposed",
- "caption":"Two product values, one raw material cost, and a $2.52 margin left over. Oil is now the larger of the two products, which is why an energy-market afternoon moves an agricultural margin.",
- "source":"CBOT Dec meal $321.60/st and Nov beans 1216¾ c/bu (18 Aug 2026 settlements); Dec oil derived from Monday's 70.96 c/lb and Tuesday's ~2.5% decline.",
- "steps":[{"label":"Meal value","value":7.08,"kind":"base"},
-          {"label":"Oil value","value":7.61},
-          {"label":"Bean cost","value":-12.17},
-          {"label":"Board crush","kind":"total"}]}
-```
-
-### The oil share, and why it inverted
-
-Look at what sits inside that $14.69 of product value. Oil is $7.61 of it — **51.8%**. Meal is the smaller half.
-
-For most of the twentieth century this was the other way round. Meal was the point of a soybean and oil was the by-product that had to be disposed of; meal routinely ran two-thirds of the value. Biofuel demand inverted it. Renewable diesel capacity built through the 2020s turned soybean oil into a feedstock competing with a fuel, and a feedstock market prices off the fuel it displaces, not off the food it used to be.
-
-The consequence for a crusher is structural, not cyclical. **More than half of what he sells is priced by regulation.** A renewable volume obligation, a tax credit, a change in what counts as a qualifying feedstock — each of these is worth more to his margin than a change in the soybean crop.
-
-Tuesday made the point cleanly. Beans finished unchanged. The crusher's margin did not.
-
-### How the trade gets put on
-
-A crush trader does not quote three legs. He quotes one number.
-
-> **CRUSHER:** Where's December board crush?
-> **BROKER:** Two fifty two, two fifty five.
-> **CRUSHER:** I'll pay two fifty three for two hundred.
-> **BROKER:** Done. Two hundred at two fifty three. Long beans, short meal, short oil. Ten, eleven, nine.
-
-Two things happened there.
-
-**One price for three contracts.** "Buying the crush" means buying bean futures and selling meal and oil futures against them. That position gains when the margin narrows — which is exactly the point, because the plant's physical business gains when the margin is wide. The paper is the mirror image of the plant, so the margin stops moving. The crusher has fixed $2.53 on 200 lots of throughput.
-
-**The 10-11-9 ratio is not a convention.** It is the seed. Ten bean contracts are 50,000 bu. That much crushes into 50,000 × 44 lb = 2.2 m lb of meal, which is 1,100 short tons, which is exactly **eleven** meal contracts. And 50,000 × 11 lb = 550,000 lb of oil, which is 9.17 oil contracts — rounded to **nine**. The ratio falls out of the physical yield and nothing else.
-
-### Board crush is not plant crush
-
-Here is the distinction that separates a trainee from a crush trader, and it is the reason a screen can mislead an entire market.
-
-That $2.52 is a paper number assembled from three futures prices. **No plant transacts at any of them.**
-
-A plant buys beans at its own gate, from farmers and elevators, at futures plus or minus a differential. It sells meal to a feed mill two hundred miles away, at futures plus or minus a differential. It sells oil to a refiner on the same basis. So:
-
-> **Plant crush = board crush ± bean basis ± meal basis ± oil basis − conversion cost**
-
-And conversion cost is real money. Natural gas to run the dryers and the desolventiser, electricity, hexane, labour, maintenance, and a depreciation charge if you are being honest about it. Variable cost at a modern plant is roughly **35–50 c/bu**.
-
-Take the friendly end of that range. $2.52 of board crush minus 40¢ is **$2.12**, before a single basis number has been added. Add a typical set of gate and destination bases — a bean basis over the board, a meal basis under it — and the number a plant actually books can be a dollar below what the screen shows.
-
-### Why a wide margin does not get competed away
-
-Standard economics says a $2 margin attracts entry until it disappears. In crush it does not, and the reason is physical.
-
-**You cannot make crushing capacity this week.** A new plant is two to three years of permitting and construction and hundreds of millions of dollars. Inside a marketing year, installed capacity is a hard constraint. When margins are wide, every plant is already running flat out, the constraint binds, and a binding constraint holds a margin open.
-
-So what adjusts?
-
-**The bean basis.** Plants bid harder for cash beans at the gate to keep the line full, because an idle hour of capacity in a wide-margin environment is the most expensive thing in the business. Origination teams push the local bid up, and up again, competing against each other and against the export elevator.
-
-The result is the single most common misreading of this market. **The board crush stays fat on the screen while the plant crush quietly compresses.** The margin is real, but it does not all end up with the crusher. A meaningful share of it is transferred to the farmer, in the basis, one truckload at a time. Anyone who models crusher profitability off the board number and no basis assumption will overstate it, sometimes badly.
-
-### What actually happens when the crush goes negative
-
-The textbook answer is that the plant shuts. The plant usually does not, and understanding why is understanding where the optionality really sits.
-
-Two reasons.
-
-**Stopping is expensive and slow.** A crush line does not idle for an afternoon. Shutting down and restarting costs money and days, and the fixed cost base does not stop.
-
-**The meal is already sold.** Most of the meal that will come out of next month's beans has been sold forward at fixed prices to feed mills that are counting on it. A plant that stops does not simply stop losing money — it has to go into the market and *buy* meal to cover its own sales, in exactly the environment where every other plant is cutting runs and meal is tightening.
-
-So the plant is not choosing between running and not running. It is choosing between a negative margin and a negative margin plus a short meal book in a rising market.
-
-The real option is therefore not on the board crush at all. **It is on the variable margin, over cash costs, on the volume that is not already committed.** The committed volume is not a decision; it is a delivery obligation, and the only question there is make-or-buy. The decision lives in the uncommitted remainder, and it is decided against cash prices and cash costs — never against the screen.
-
-And the system repairs itself. Run rates fall, meal supply tightens, meal rallies, the crush widens, the plants come back. Which is why a deeply negative board crush is more often a reason to own meal than a reason to sell it.
-
-### The leverage, in one number
-
-Finish with the ratio that makes all of this matter.
-
-The margin is $2.52. It sits on $14.69 of gross product value. The crush is a **17% margin on the products, built from a position with five to six times its own size in gross exposure.**
-
-So a 1% move in the product complex is worth about 15¢ — roughly **6% of the entire margin**.
-
-Tuesday's oil move on its own makes the point. A 2.5% decline is about 1.77 c/lb, and 1.77 × 0.11 = **19½ c/bu**. Nineteen and a half cents against a $2.52 margin is close to **8% of a crusher's economics**, delivered in one afternoon, by a market that is not soybeans and never was.
-
-```chart
-{"type":"bar","unit":"$/bu impact on the crush",
- "title":"What a 10% move in each leg is worth",
- "x":["Meal +10%","Oil +10%","Beans +10%"],
- "series":[{"name":"Change in board crush","values":[0.71,0.76,-1.22]}],
- "caption":"The bean leg is the biggest single exposure — 1.6 times either product — which is why it is the leg to hedge first even though oil is the volatile one. Volatility matters, but only multiplied by notional.",
- "source":"Computed from Tuesday's board crush legs: meal $321.60/st, oil ≈69.2 c/lb, beans 1216¾ c/bu."}
-```
-
-Three things to keep. The two multipliers, 0.022 and 0.11, because the whole market runs through them. That the board crush is a quote and not a margin, with three bases and a conversion cost standing between the two. And that more than half of a soybean's value is now a fuel — so a crusher who watches only beans is watching the smallest of his three prices.