Start with what "investing in commodities" actually means, because it is rarely what a newcomer pictures. Almost no institutional or retail commodity exposure involves owning the physical good, nobody's pension fund has a warehouse of soybeans. The standard access route is a futures-based index (S&P GSCI, the Bloomberg Commodity Index, or a fund/ETF that tracks one of them): a continuously rolled position in near-dated futures contracts, collateralized in cash or government bills, never taking physical delivery. Understanding the return of that position, not the spot commodity price alone, is the heart of this section.
How do the three commodity index levels fit together?
Commodity index providers publish (and you should think of your own exposure in terms of) three parallel index levels, and the gap between them is not noise, it is the whole story:
- Spot Index: tracks only the change in the nearest/spot commodity price. No futures, no collateral, just "what did the physical price do."
- Excess Return (ER) Index: tracks a continuously-rolled long futures position, uncollateralized. This is what an investor actually captures by holding the futures contract rather than the physical spot good, roll after roll.
- Total Return (TR) Index: the ER index plus the interest earned on the (typically 100%) cash/T-bill collateral backing the position. This is the return an actual fully-funded index investor receives.
From these three levels, the module's central identity falls out directly:
Roll Yield (over a period) = Excess-Return-Index return − Spot-Index return
and
Total Return = Excess Return + Collateral Return = (Spot Return + Roll Yield) + Collateral Return
Advertisement
Worked example: a month in contango
Work it with clean numbers first, in an ordinary month with no crisis in it (Python-verified in Worked Example 1): suppose over one month the Spot Index rises 2.10% while the ER Index rises only 0.35%. The roll yield for that month is 0.35% − 2.10% = −1.75%, the futures market was costing the rolled position 1.75 percentage points relative to just holding the spot good, because (as the contango and backwardation page explains) the curve was in contango. Add a collateral return of a 1-month T-bill at 5.20% annualized (≈0.4333% for the month), and the Total Return is 0.35% + 0.4333% = 0.7833%, which reconciles exactly against Spot + Roll + Collateral = 2.10% + (−1.75%) + 0.4333% = 0.7833%. Every number in that paragraph is reproduced from first principles in Worked Example 1, check it yourself once, because the identity (not the specific numbers) is the tool you carry forward.
Why can a commodity fund lag the spot price?
Three practical lessons sit inside that one piece of arithmetic. First, "commodities are up" and "your commodity fund is up" are different claims, a headline about the spot oil price tells you almost nothing about what a rolled futures position actually earned that month, because the roll yield can run strongly against the spot move (this is precisely why long-only crude-oil ETFs have, in real episodes, badly lagged the spot price of oil, the ETF holds the futures, not the barrel, and the roll yield ate the difference). Second, collateral yield is not a rounding error, in a high-short-rate environment (a 5%+ T-bill world), the collateral leg alone can be several percentage points a year, which is why "commodities" performed very differently as an asset class in the near-zero-rate 2010s versus a higher-rate regime; a commodity index's total return has a bond-like component riding underneath its commodity-like component, and conflating the two misreads what drove the number. Third, and this sets up contango and backwardation and GSCI vs BCOM: the sign of the roll yield is not random, it is set by whether the futures curve is in contango or backwardation, and that shape has real economic causes worth understanding rather than memorizing.