Enhanced Commodity Index Roll

Thesis (edge)

A traditional commodity index does something remarkably naive. It holds the nearest contract, and on a set of pre-announced days each month it sells that contract and buys the next one. Everybody knows the schedule. It is published in advance. And the index is very large.

Two separate problems follow. The first is that the front of the curve is often the worst place to be. In a market in contango, where deferred contracts cost more than nearby ones, holding the front contract means you repeatedly sell something cheap and buy something expensive. That is a steady, grinding cost. Between 2009 and 2014, this alone destroyed a large part of the return of long-only commodity index investors, even in periods when spot prices were flat or rising.

The second problem is that trading on an announced schedule is an invitation. Other traders know the index must sell the front and buy the next contract on those specific days, and they can position ahead of it. The index pays for that.

The enhanced version fixes both. It picks the contract on the curve with the best roll characteristics rather than blindly holding the front, and it rolls on a schedule that is spread out and not perfectly predictable.

Where it works (regimes)

The contract selection part matters most in contango, which is the more common state for storable commodities like crude oil and natural gas. When the curve is steeply upward sloping, moving further out along the curve avoids the steepest part of the decay and can save several percent a year.

The roll timing part matters most when index roll flows are large relative to market liquidity, which is more true in the smaller agricultural markets than in crude oil.

Neither part helps when commodity prices themselves are falling. This is important and often glossed over. An enhanced index is still a long-only commodity position. It loses money in a commodity bear market, just slightly less than the standard index does.

Signals

• For each commodity, compute the implied annualized roll yield of holding each available contract month. This is the slope of the curve between that contract and the one after it, scaled by time.
• Choose the contract with the best implied roll yield, subject to liquidity. In practice this means going further out the curve in contango and staying near the front in backwardation.
• Apply a liquidity floor: reject any contract whose open interest or volume is below a threshold, no matter how attractive its roll yield looks.
• For roll timing, either spread the roll evenly across a window of days, or shift the window away from the standard index roll days, or both.

Portfolio construction

Keep the commodity weights the same as the benchmark unless you intend to take active weight bets too. That keeps the comparison clean: the enhancement is purely about which contract you hold and when you roll it, not about which commodities you like.

If you do want to overweight commodities with favourable curves, that is a different strategy and belongs in the carry family rather than the index enhancement family. Mixing the two makes it impossible to know which part is working.

Risk model

The main risk relative to the benchmark is tracking error. You will hold different contracts, so on any given day you will not match the index. In a fast-moving market the deferred contract you hold may move differently from the front contract the index holds, and clients who expect index-like behaviour will notice.

The second risk is liquidity. The further out you go, the thinner it gets. In a stressed market, deferred contracts can become effectively untradeable, and the very moment you want to reduce risk is the moment you cannot.

Costs & implementation

The point of the strategy is to save costs, so measuring them properly is the whole game. Deferred months have wider spreads. If you save 3 percent a year in roll yield and give back 1 percent in worse execution, your net gain is 2 percent, and that is still excellent, but you must know both numbers.

Spreading the roll over multiple days reduces market impact and reduces the chance of being picked off on a single bad day. It also introduces some randomness in your fills, which is exactly the point.

Capacity is good in the large energy and metals markets and constrained in agriculture. If you run enough money, you become the flow that others front-run, and the strategy starts to work against itself.

Failure modes

• Being seduced by roll yield into holding contracts that are too far out and too illiquid.
• Forgetting that you are still fully long commodities and the enhancement is only a cost saving.
• Tracking error surprising clients who signed up for an index-like product.
• The roll timing edge decaying as more capital chases it. It is a crowded and well-documented pattern.
• Backtesting on deferred contract prices that were quoted but never really tradeable.

Our Notes & Suggestions

This is one of the more honest strategies in the whole space, because the source of the edge is not mysterious. A very large, price-insensitive, calendar-driven buyer exists in the market, and you are simply choosing not to be that buyer. Everything you gain comes from avoiding a known cost rather than from predicting anything.

Report your results as an excess return over the standard index, not as a total return. The total return is dominated by commodity prices, which you have no view on, and it will make a good enhancement look terrible in a bad commodity year.

Expect the timing part of the edge to decay. The contract selection part is more durable, because it is tied to the real economics of storage rather than to somebody else's predictable behaviour.