Nominal Spread 3.5%, Net Margin 0.35%: The Five Layers That Eat It

"Our margin on spot-plus-futures trading is not high" — this is the sentence trading and manufacturing companies most often hand to an advisor. But it is not a diagnosis. It is a symptom. From quote to final cash, the nominal spread passes through five layers of deduction, and each layer takes its cut in a different way.

The first reaction in most companies is to squeeze suppliers, or to ask customers for a higher price. Both are useful. But if the main leak is not in commercial negotiation, negotiation recovers only part of it. The rest is structural leakage — and a structural leak cannot be talked away.

Start with one real account: how a 3.5% spread becomes 0.35%

Here is a typical trading business, with every layer put back the way it happened:

The nominal spread between the selling price and the buying price is 3.5%. That number looks normal, and it is indeed the figure many companies report. But take the whole business through:

• Basis P&L −1.2%: pricing happened after sales had locked the price and before purchasing had locked its own. The basis weakened in that window, and the spread was eaten directly. On paper "the spread is 3.5%"; what actually lands in the account is 2.3%.
• Timing gap −0.8%: the sales side had already fixed its price while the purchasing side was still waiting for supplier quotes. The price difference in between is naked exposure, and it does not disappear on its own.
• Holding cost −0.6%: the goods sat in hand for three months. Capital tied up, storage, insurance, plus the risk of a price drop — all of it paid out of the 3.5%.
• Hedge cost −0.4%: to hedge that 0.8% of mismatch exposure, options were used. Premium amortisation plus basis mismatch took 0.4% in total.
• Trading cost −0.15%: pricing, opening and closing positions, settlement. Small on a single trade, but clearly significant once volume rises.

What is left at the end is a net margin of 0.35%. On RMB 1 billion of annual revenue that is RMB 3.5 million — a full year of profit, eaten by one business line, while the gross margin on the statement still shows 3.5%.

The most important information in that chart is not the five numbers. It is this: they sit in five different departments, and each one needs a different action to repair it. That is why "everyone works a bit harder" solves nothing — nobody knows how much their own layer took.

Five leak layers, taken apart one by one

Layer 1 · Basis P&L: the biggest leak, and the one most often ignored

The core of the basis problem is the choice of pricing moment. Many trading companies have turned pricing into a business in itself: the customer prices whenever it wants to, and the moment is decided by the customer's appetite and how tight physical supply is — not by the company.

The result is this: the company is in effect carrying basis risk on behalf of the customer. The customer needs the goods urgently at a low price level, the company accepts the pricing moment to protect the deal, and when the basis weakens the profit is eaten, while when it strengthens the gain often does not stay either.

Identification is simple. Sort every pricing event in the past 12 months by the basis percentile at the moment of pricing, and see whether the company's pricing moments are systematically outside the favourable band. If most pricing lands in the unfavourable zone, that is not bad luck. It is mechanism.

There are three levers, ranked by how fast they work: first, unify the pricing window (so customers cannot price and take delivery whenever they please); second, change the pricing clause from an absolute price to "basis plus pricing window", so both sides share the basis; third, agree upper and lower bounds on the pricing range in the contract, locking out the extreme cases in advance.

Layer 2 · Timing gap: sales locks early, purchasing locks late

This layer is the easiest to identify: look at how many days separate the "sell lock date" from the "buy lock date" on a single order.

If sales locks on day T and purchasing locks on day T+60, those 60 days are a pure exposure window. No matter how well you hedge afterwards, the loss inside that window is already determined. More awkward, this exposure usually appears in no risk report at all, because it sits in neither the futures position nor the contract ledger.

The fix is to make sales orders and purchasing plans roll on the same rhythm: once the sales plan is clear, purchasing and hedging start together, instead of reacting only after supplier quotes come back. Get this right and the second layer of the waterfall above goes straight to zero.

Layer 3 · Holding cost: the layer everyone treats as "no way around it"

Capital tied up, storage, insurance, price-decline risk — these are indeed rigid costs, but they can be priced.

Concretely: combine "cost of capital + storage and logistics cost + risk cost" into one composite holding cost rate, annualised, and then add it into the quote.

Many companies quote on "commodity cost + target profit" and never compute the cost of the holding period at all. The result: the bigger the volume and the higher the inventory, the better the margin looks on paper, while in reality the business is trading more capital for less profit.

Layer 4 · Hedge cost: it is not hedging that is expensive, it is how hedging is used

This layer is the one most easily blamed. The cost of hedging is not only the premium. It has three parts:

• Explicit cost: option premiums and transaction commissions;
• Hidden cost: hedge-effectiveness loss caused by basis mismatch — often far larger than the premium;
• Capital cost: the interest on margin tied up.

The real problem usually sits in how the instruments are used: every exposure hedged with the same instrument, every timing handled with the same strategy. So the part that futures could have covered cheaply ends up wearing an options overlay, and the part that natural hedging could have absorbed ends up wearing a financial instrument.

The direction is layered instrument use: futures for the certain, naturally hedgeable part; options as a floor under the uncertain but important part; and for what is genuinely unimportant, or where cost exceeds benefit, honestly choose not to hedge and say why. "Not hedging" is also a reasoned decision.

Layer 5 · Trading cost and scale diseconomy: more volume is not better

Commission and slippage on a single trade look negligible. But once order size grows and the individual trade shrinks, it becomes non-linear: process complexity rises, manual error rates rise, output per person falls, and in the end scale went up and profit did not.

Here is a counter-intuitive but important conclusion: expanding volume before the process is fixed amplifies losses rather than amplifies profit. Get a single trade to run cleanly and bring the error rate down first. Then talk about scale.

Put the five layers side by side and each one has the same shape: a symptom the business notices, a cause that is usually not where the symptom appears, and a lever that is specific to that layer. Diagnosing by symptom alone is what sends teams to the wrong department.

Four improvement levers, ranked by how fast they work

Turning the analysis above into action, we usually recommend this order:

1. Do the spread waterfall decomposition first (2–4 weeks). Reconstruct the last 6–12 months of business across the five leak layers. This is the foundation for everything that follows, and the cost is very low — usually one round of data from finance and one from the business side.
2. Lock the basis terms (1–2 months). Move pricing from an absolute price to basis plus a pricing window, and agree upper and lower bounds on the pricing range. This is the fastest layer to improve, because it depends on no system build at all.
3. Put sales and purchasing on the same rhythm (3–6 months). Establish an order-driven rolling purchasing and hedging mechanism that removes the timing gap window. This layer needs cross-department coordination. It is the slowest and also the most valuable.
4. Price by turnover tier (continuous). Put the holding cost into the quote and let the price mechanism adjust turnover, instead of chasing it with administrative instructions.

A self-check you can run today

1. What is your actual net margin over the last 12 months? Note the question: the net margin after basis P&L, holding cost and hedge cost — not the gross margin on the statement.
2. Where is your pricing-moment distribution? If customers can price and take delivery whenever they want, you are in effect carrying basis risk for them — is that written into the contract?
3. On a typical order, how many days pass between the sell lock and the buy lock? Can you see that window in your risk report?

If you cannot answer the first of these three questions, the problem is most likely not that your margin is low. It is that you are not yet looking at your own business at true cost. That is why we suggest starting from a diagnosis rather than from a price increase.

Data sources and notes

The spread waterfall decomposition framework and the five-layer leak structure used in this article are a standard analysis framework in RMC's commodity risk management diagnostics for trading and manufacturing companies. The illustrative figures are anonymised constructions of typical situations, used to demonstrate the decomposition method; they do not represent the operating data of any specific company. Companies can reconstruct the same view from their own accounts using the same method. This article does not constitute a recommendation for or an evaluation of any specific software product.