Roll Yield & Commodity Indices: Why Buy-and-Hold Leaks Money

You did the hard yards in lessons 1 and 2. You can read a futures contract and its basis (spot minus futures), and you can build the curve out of cost of carry — financing plus storage minus convenience yield — which is exactly why the curve slopes up in contango and down in backwardation. Good. Hold that thought, because we’re about to spend it.

Here’s the uncomfortable truth this lesson exists to deliver: you cannot hold a futures position forever. Contracts expire. The June crude contract does not quietly become “crude forever” — on a fixed date it stops trading, and if you’re still holding it the exchange would very much like you to take delivery of 1,000 barrels of West Texas Intermediate at a tank farm in Cushing, Oklahoma. You do not want that. So to keep continuous exposure, you roll: sell the expiring contract and buy a later-dated one. And that roll, whose price is already baked into the curve shape you learned last time, quietly makes or loses money on every single turn.

This is the single most misunderstood thing in commodity investing. People buy a “long oil” product, watch spot oil rise 30%, and somehow finish the year down. They didn’t get robbed. They got rolled. Let’s make sure that never confuses you again.

Why you have to roll

Think of a magazine subscription. You love the magazine; you intend to read it forever. But the subscription you bought covers this year only. When it lapses you have to renew — and the renewal price is whatever the publisher charges next year, not the price you locked last year. Futures are the same. A single contract is a one-year (or one-month) subscription to price exposure. To “keep reading,” you renew into the next contract at its price, not the price of the one expiring.

Definition. Rolling a futures position means closing the contract nearest to expiry (the front month) and simultaneously opening a position in a later-dated contract (the next month out), so your exposure is continuous and you never get within delivery range. A passive long does this on a schedule, every roll period, for as long as they want exposure.

The mechanics for a long are dead simple: you sell the front contract you hold and buy the next contract. The catch is the prices you sell and buy at, which are two different points on the curve.

Let’s make it concrete. Say it’s late May, you’re long one crude contract, and the June contract is about to enter its delivery window. You roll into September.

You sold at $103 and re-bought at $106. Same barrels of exposure, but the renewal cost you $3 per barrel of gap — and notice you did this without spot oil moving a cent. The price difference came entirely from the shape of the curve, which is the cost-of-carry wedge from lesson 2.

When to use it

You roll whenever you want exposure that outlasts a single contract — which is essentially all passive, long-term commodity investing. If your horizon fits inside one contract’s life (a one-off hedge that expires when your risk does), you don’t roll at all and this whole lesson is someone else’s problem. The moment your horizon exceeds the contract’s, rolling — and its cost — is unavoidable.

Roll yield: the hidden P&L

Here’s the part that pays the rent. The total profit or loss on a rolled futures position splits cleanly into two pieces: the spot move (did the underlying commodity actually get more or less expensive?) and the roll yield (what did the act of rolling cost or earn, independent of spot?). They’re different, they can point in opposite directions, and confusing them is how people lose money while being “right” about oil.

Definition. Roll yield is the return generated by the roll itself — the gain or loss from selling the expiring contract and buying the next one — holding the spot price constant. It is governed entirely by curve shape:

• Contango (next contract dearer than front): you sell low, buy high every roll → negative roll yield, a persistent drag that bleeds even when spot is dead flat.
• Backwardation (next contract cheaper than front): you sell high, buy low every roll → positive roll yield, a tailwind that pays you to hold.

A clean way to approximate it per roll is the percentage gap between the two contracts you trade:

$\text{roll yield} \approx \frac{F_{\text{front}} - F_{\text{next}}}{F_{\text{next}}}$

Watch the sign. When the front is below the next (contango), the numerator is negative → negative roll yield. When the front is above the next (backwardation), it’s positive. Play with the curve and the roll legs here:

Now two fully worked examples — one curve of each shape.

Contango roll (the bleed). Front June at $103, next September at $106.

You’re down ~2.8% on this roll before spot oil does anything. Do that monthly and the annualized drag is brutal.

Backwardation roll (the tailwind). Front at $100, next cheaper at $97.

Same mechanical roll, opposite sign, because the curve sloped the other way. You sold high and re-bought low, pocketing ~3.1% just for rolling.

When to use it

Roll yield is the lens you use to judge any passive long-only commodity position. Before you buy continuous exposure, read the curve: a persistently contango’d market (think much of the time in crude, natural gas, gold) means you’re paying to hold, and you need a spot move big enough to clear that toll. A backwardated market is paying you to hold — historically the regime where long-only commodity strategies have actually earned their keep. Don’t pick a commodity product on spot view alone; price the roll first.

Total return = spot + roll + collateral

Now zoom out from one roll to the whole strategy. Because a futures position only needs a sliver of margin, a fully-collateralized commodity investor parks the rest of the cash — essentially the full notional — in safe short-term Treasuries. That cash earns interest. So the honest accounting of what you actually pocket has three ingredients, not one.

Think of it like a food-delivery business. Your spot return is how much the meals themselves went up or down in price. Your roll yield is the courier’s surcharge — sometimes a fee, sometimes a discount — for keeping the kitchen open continuously. And your collateral return is the interest your float earns sitting in the bank while you wait. The headline (“oil went up!”) is only the first ingredient.

Definition. The total return of a fully-collateralized commodity index decomposes as:

$\text{Total return} = \underbrace{\text{spot return}}_{\Delta\,\text{front price}} + \underbrace{\text{roll yield}}_{\text{curve shape}} + \underbrace{\text{collateral return}}_{\text{T-bill interest}}$

• Spot return — the change in the front-contract price; the “did the commodity get more expensive” part.
• Roll yield — the contango drag (−) or backwardation gain (+) from lesson logic above.
• Collateral return — risk-free interest (T-bills) on the cash backing the futures, because futures consume little margin and leave most of the notional free to earn yield.

Here’s the stacked decomposition — see how the three bars build (or cancel) into the total:

Worked example — the year roll drag dominates. Spot rose, the curve was steep contango, and bills paid a healthy yield.

Spot oil went up 5%, yet the roll bled −7%. Without the +4% collateral cushion the investor would have lost money on rising oil. The headline number (+2%) bears almost no resemblance to the spot story (+5%). This is the wedge that wrecks naive “long oil” bets.

Contrast year — backwardation tailwind, low rates. Now flip the curve and drop rates.

Same spot-ish move (+3%), but the curve paid you +6% to roll and you finished +10%. Identical-looking commodities, wildly different total returns — entirely because of curve shape and rates, not the spot view.

Quick refresher on the curve shapes driving all of this — toggle between them:

When to use it

Use the three-part decomposition any time you evaluate a commodity fund, ETF, or index swap. Don’t judge it on spot or even spot-plus-roll alone — in a high-rate environment the collateral leg can be the difference between a positive and negative year, and in a steep-contango market the roll leg can quietly erase a good spot call. When you read a fund’s fact sheet, mentally split its return into these three and ask which one is actually carrying the strategy.

Commodity indices: GSCI vs BCOM

So how do real products package all this? Through commodity indices — rules-based baskets of rolled, collateralized futures. The two benchmarks everyone quotes are the S&P GSCI and the Bloomberg Commodity Index (BCOM), and the difference between them is entirely about weighting philosophy.

Picture two grocery baskets. The GSCI fills its basket in proportion to how much of each commodity the world produces — and since the planet produces a staggering amount of oil and gas by dollar value, the GSCI basket is mostly energy. It’s a faithful mirror of the global production economy, which also means it lives or dies on crude. The BCOM instead writes a rule: no single commodity or sector may exceed a cap, so it deliberately tops up metals and agriculture to keep any one bet from swamping the basket. It’s the “balanced diet” index.

Definitions.

• S&P GSCI — world-production-weighted; energy (especially crude) dominates. Maximal commodity-beta, minimal diversification, very sensitive to the oil curve.
• Bloomberg Commodity Index (BCOM) — capped weights per commodity and sector, forcing balance across energy, industrial metals, precious metals, grains, softs and livestock.

What they share matters just as much: both are long-only, both are front-loaded (they hold near-dated contracts), both are fully collateralized, and both roll on a fixed published schedule. That last point is the kicker — because both roll mechanically on a calendar, both suffer contango drag whenever the front of the curve is in contango. Their diversification differs; their exposure to the roll wedge does not disappear.

This is exactly why “enhanced” or second-generation indices exist. Instead of always rolling into the next listed contract, they try to roll into whichever part of the curve is cheapest to hold — picking a contract further out where the curve is flatter, or optimizing the roll date — specifically to reduce the contango drag you now understand cold. They don’t abolish roll yield; they just stop bleeding from the worst point on the curve.

When to use it

Reach for a broad commodity index when you want inflation hedging or portfolio diversification — commodities historically zig when stocks and bonds zag, and they tend to rise with the inflation that hurts nominal bonds. Pick GSCI-style exposure if you want a high-octane, energy-led commodity beta; pick BCOM-style if you want a steadier, more diversified ride. But go in eyes open to roll cost: in a persistently contango’d world, the index can lag spot for years, so size the position for the total return — spot plus roll plus collateral — not the commodity headline. If roll drag is your main worry, an enhanced/second-generation index is the targeted fix.

Putting it together

Step back and the whole chain clicks. A passive commodity long must roll, because contracts expire and nobody wants 1,000 barrels delivered to a tank farm. That roll trades at two points on the curve — sell the front, buy the next — so its sign is decided by curve shape: a drag in contango, a tailwind in backwardation. Stack that roll yield on top of the spot return and the collateral (T-bill) return, and you get the total return that a commodity index actually delivers — which is emphatically not the spot commodity’s move. The GSCI mirrors world production (energy-heavy); BCOM caps for balance; both still pay the contango toll, and enhanced indices try to roll smarter. The headline “oil went up” was never the whole story — the curve and the collateral wrote the rest.

Key Takeaways