Estimating Early-Expiry Volatility for Commodity Futures Options

Summary

The document asks how to estimate short-dated implied volatility for options on commodity futures when listed expiries are sparse. Its soybean example contrasts contracts with clearly observable weekly or short-dated option expiries against intermediate contracts that have only one or no early-expiry observations. The central concern is that options tied to different delivery contracts may not share the same volatility behavior, especially when seasonal effects create differences between contracts.

Several possible approaches are raised: borrow volatility from another contract, extrapolate listed total variance through time, or fit a model such as Samuelson or SABR to describe the decay. The text does not supply an accepted solution, calibration results, or a comparison of model performance; it is a research question rather than evidence that one method is superior. Any practical estimate would need to account for contract-specific seasonality and the limited observations available at the short end.

Key ideas

• Listed option expiries may leave gaps in the short end of commodity futures volatility surfaces.
• Using volatility from another delivery contract risks overlooking seasonal differences between contracts.
• Extrapolating total variance is one possible way to estimate missing early expiries.
• A time-decay model such as Samuelson or SABR is another proposed approach.
• The document raises alternatives but does not establish a preferred method or validate one empirically.

Tags

# Early expiry volatilities in commodity futures markets


# Early expiry volatilities in commodity futures markets












I'm trying to figure out what the short end of the volatility term structure for options on commodity futures looks like.

A concrete example: Soybean (ZS / S A Cmmdty)

- X25 -> short term weekly options (short end is well defined)

- F26 -> has one monthly option so only two points in the term structure if we discard all the vols that look at the X25 underlying contract

- All the mid term contracts -> no early expiry options

- X26 -> Short dated new crop options -> very well defined term structure

For me it is clear that the X25 and X26 contracts are well defined and should get their own vol surfaces, no struggles there. My issues lie with the intermediate contracts, those with 1/0 early expiry options: how do I get the vol for OTC early expiry options on these contracts? The easy, spot market style answer is to price with the vols from the other contracts, treating them all as the same underlying, but this feels like it completely disregards the dislocation between contracts (especially wrong for highly seasonal commodities). Another simple approach would be to extrapolate linearly (in total variance) and decay the listed vols in time. A somewhat better approach would be to model this decay in time with a model, such as Samuelson (or something like SABR?).

What is the correct/accepted approach for this? What are the pros and cons?

Thanks!

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.