Derivatives — forwards, futures, swaps

M0.5 gave you the textbook line: a derivative is a contract whose value derives from an underlying. True, and almost useless — it tells you what a derivative is named after, not what it is.

Here’s the version that actually helps: a derivative is an agreement, made now, about a transaction that happens later. The thing being traded is the agreement itself. Almost nobody who trades an oil future wants the barrels.

Everything else — margin, clearing, expiry, leverage — falls out of that one idea plus the practical problems of making such an agreement stick. So let’s build one from scratch rather than memorise the vocabulary.

Start with the problem, not the product

A farmer will harvest 100 tonnes of wheat in September. He doesn’t know what it’ll fetch. If the price collapses he may not cover his costs.

A miller will need 100 tonnes of wheat in September. She doesn’t know what she’ll pay. If the price spikes her margins vanish.

They have exactly opposite risks. Neither of them wants a gamble — they each want to know their number so they can plan. So in March they agree: in September, the farmer delivers 100 tonnes and the miller pays £200 a tonne. No money changes hands today. That agreement is a forward contract, and it’s the atom of the entire derivatives world.

   September spot     farmer      miller       who’d rather have
   price of wheat     receives    pays         had no contract?
   ──────────────────────────────────────────────────────────────
      £150/t            £200        £200        the miller (−£50/t)
      £200/t            £200        £200        neither
      £250/t            £200        £200        the farmer (−£50/t)

Read that table carefully, because it contains the point most explanations skip. Neither of them “won”. One of them will always look, in hindsight, like they’d have done better without it — and that is not a flaw, it’s the cost of the thing they actually bought, which was certainty.

Notice too what happened to the risk: it didn’t disappear. It moved. Price risk left two people who didn’t want it and landed on… each other, cancelled out. That’s the whole purpose of derivatives, and it’s worth holding onto when the machinery gets complicated.

And notice the contract itself now has a value that moves. In September with wheat at £250, the miller’s right to buy at £200 is worth £50 a tonne to her. The contract started life worth nothing to either side and is now worth something — an amount derived from the wheat price. Hence the name.

Two things wrong with a handshake

The forward works. It’s also fragile in two specific ways, and every piece of futures machinery exists to fix one of them.

1. Counterparty risk. September arrives, wheat is £250, and the farmer can sell to anyone for £250. Does he honour a £200 deal? If he refuses, or goes bust, the miller’s hedge evaporates precisely in the state of the world where she needed it. A hedge that fails in the bad state is not a hedge.

2. No exit. In June the miller’s plant burns down. She doesn’t need wheat any more, but she’s still contractually obliged to buy it. Her only ways out are to persuade the farmer to tear it up, or find someone to take her place — which needs the farmer’s agreement, because he’s relying on her creditworthiness, not a stranger’s.

Both flaws come from the same root: it’s a bilateral, bespoke contract between two named people.

The future: an exchange fixing both

A future is a forward with three modifications, each one solving something above.

1. Standardize it. The exchange, not the counterparties, defines the contract: 100 tonnes, this grade of wheat, this delivery month, this warehouse. Only the price is negotiable. Now every September wheat contract is identical to every other — fungible, exactly like shares in M0.5 — so it can trade in an order book with all the M0.2 machinery.

That solves the exit problem in an elegant way: you don’t unwind your contract, you take the opposite position. Long one September contract and short one September contract cancel to nothing, and the clearing house nets them out of existence. This is offsetting, and it’s how the overwhelming majority of futures end — closed out before expiry, no wheat ever moved.

2. Put a clearing house in the middle. Once the trade is agreed, the CCP from M0.6 steps between the two sides by novation: it becomes the buyer to every seller and the seller to every buyer.

   BILATERAL FORWARD                    CLEARED FUTURE

   farmer ◀──────────────▶ miller       farmer ◀──▶ CCP ◀──▶ miller

   each relies on the other             each relies only on the CCP
   the other defaults → hedge dies      CCP guarantees both sides
   exit needs the other’s consent       exit = trade the opposite

Counterparty risk solved — or rather, concentrated into one institution built to survive it, which is a different thing and the reason CCPs are so heavily regulated.

3. Margin and daily mark-to-market. The clearing house has now volunteered to absorb everybody’s default risk. It is not going to do that unsecured. So it demands two things:

• Initial margin — a good-faith deposit, typically a few percent of the contract’s value, posted before you trade.
• Variation margin — every evening the contract is repriced and the day’s profit or loss moves in actual cash between accounts. Winners are paid, losers pay. If your balance falls below the maintenance margin, you get a margin call: top it up by morning or be closed out.

Why settle daily instead of at expiry? Because it stops losses quietly accumulating into a number the loser can’t pay. The clearing house never lets anyone owe more than one day’s move. That single design decision is why futures markets survive events that destroy OTC counterparties.

What margin actually feels like

Concrete numbers, because this is where intuition usually fails.

A wheat future covers 100 tonnes. Wheat is £200/t, so one contract represents £20,000 of wheat — its notional. Initial margin is £1,200 (6%), maintenance margin £900. You buy one contract: you are long, and you profit if wheat rises.

         settle   move    P&L on    margin
   day   price    per t   100 t     balance   what happens
   ─────────────────────────────────────────────────────────────────
    0    200.00      —        —      1,200    you post initial margin
    1    203.00   +3.00    +£300     1,500    £300 paid INTO your account
    2    201.50   −1.50    −£150     1,350    £150 taken out
    3    196.00   −5.50    −£550       800    below £900 → MARGIN CALL
    3′      —        —        —      1,200    you wire £400 by morning
    4    199.00   +3.00    +£300     1,500

Sit with row 3. Wheat fell 2%, from £200 to £196 — an unremarkable day — and you had to find £400 in cash the next morning or have your position liquidated. That’s leverage, and it isn’t a figure of speech:

   you control:   £20,000 of wheat
   you posted:    £ 1,200
   ratio:         ≈ 17×      → a 6% move against you wipes out the deposit

Two implications people routinely miss:

• The £300 on day 1 is real cash in your account, not a paper gain. Futures P&L is realised daily. That’s pleasant on the way up and unforgiving on the way down.
• Your deposit is not your maximum loss. Buy a share and the worst case is −100%; hold a future through a violent move and you can owe more than you ever put in.

The family in one table

                 who is obliged?     cash up front    traded      exit by
   ──────────────────────────────────────────────────────────────────────────
   forward      both sides, always      nothing       OTC         negotiate
   future       both sides, always      margin        exchange    offsetting
   swap         both sides, always      margin *      OTC/cleared offset/unwind
   option       one side only           premium       both        offset/expire

   * post-2008 reform pushed most standardized swaps into central clearing (M0.6)

Swaps sound exotic and aren’t. A swap is an agreement to exchange two streams of payments over time. The workhorse is the interest rate swap: for five years, you pay a fixed rate and receive a floating rate on an agreed notional of, say, £10m. The notional itself never changes hands — only the difference in interest each quarter. A company with a floating-rate loan that wants certainty pays fixed on a swap; the two floating legs cancel and its borrowing is synthetically fixed.

The clean way to hold it: a forward is one exchange in the future; a swap is a strip of forwards, the same logic repeated every quarter.

Options are the odd one out — the only member where one side holds a right rather than an obligation. That asymmetry changes everything about how they’re priced and used, which is why they get M0.8 to themselves.

Why anyone bothers

1. Hedging — transferring risk you don’t want. The farmer and the miller. An airline fixing jet fuel costs. An exporter locking tomorrow’s FX rate. A pension fund with fixed liabilities using rate swaps. The aim isn’t profit; it’s converting an unknown into a known so the real business can be planned.

2. Speculation — taking risk you do want, cheaply. To bet £20,000 on wheat physically you need £20,000 and somewhere to store 100 tonnes. In futures you need £1,200 and a brokerage account. This is often framed as the disreputable half, but it isn’t optional: the farmer needs someone willing to take the other side, and there aren’t always enough millers. Speculators are the counterparties that make hedging possible, and their trading is what discovers the price (M0.1).

3. Access — the reason people forget. Derivatives let you take positions that are awkward or flatly impossible in the underlying: shorting without borrowing the asset, taking a view on an interest rate or on volatility (neither of which you can hold in a warehouse), getting index exposure without buying 500 separate shares, or trading oil without a tanker.

How they hurt people

Leverage is symmetric. Seventeen times up is seventeen times down, and the margin call arrives on the worst possible morning — when the market has already moved against you and cash is hardest to raise.

Being right isn’t enough; you have to survive being right. A hedge that is entirely correct about September can still bankrupt you in June, because the derivative marks to market daily in cash while the offsetting gain in your physical business only shows up later, at delivery. That timing mismatch — right position, wrong cash flow — has killed real companies that were, technically, hedged.

The hedge only works if the underlying matches. Hedge jet fuel with crude futures, because there’s no liquid jet fuel contract, and you’ve swapped price risk for basis risk — the risk that the thing you hedged with and the thing you own stop moving together. Basis risk is the fine print on every practical hedge.

Source: John Hull, Options, Futures, and Other Derivatives, ch.1–2 (mechanics of futures markets) and ch.7 (swaps); Harris ch.2 for where derivatives sit among the asset classes. For the flavour of how specific these contracts are, open any exchange’s contract specification page (CME wheat, ICE Brent) — the obsessive detail is exactly the standardization this lesson describes.