Handling Futures Rolls in Daily Return Series

Summary

The document asks how to build a multi-year time series for a three-month futures contract while handling contract rolls. The author is concerned that prices are volatile on roll days and asks about interpolation or using calendar spreads to create roll-adjusted prices.

The response proposes focusing on daily futures price changes for variance calculations and omitting the roll day when constructing that change series. This is a narrow treatment of the problem: it does not describe a continuous adjusted price series, compare adjustment conventions, or explain how to handle roll effects for other statistics or trading signals. It also provides no empirical comparison or test. The suggestion is therefore relevant when the objective is variance from daily changes, but it does not fully answer how to build a general-purpose, roll-adjusted multi-year futures price history.

Key ideas

• The document concerns stitching a long futures history across contract rolls.
• The response recommends calculating variance from daily price changes and omitting the roll day.
• It does not specify a continuous price adjustment method or discuss interpolation and spread-based proxies.
• The suggested treatment is limited to variance calculations and is not presented with empirical validation.

Tags

# time series for futures roll


# time series for futures roll












I'm trying to build a multi year time series for a 3 month futures contract. How do I handle rolls? On the day of roll, volatility is high and I want to roll over to next contract series in a way that gives roll adjusted prices.

What sort of interpolation I can use or proxy using spreads? ANy suggestions welcome

## Answer by dm63 (score 1)

https://quant.stackexchange.com/a/33952

The Var depends only on the daily changes in futures price. So you create a series of daily changes and you simply omit the day when the roll occurs.

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.