Rounding Cointegration Strategy Inventories for Futures Contracts

Summary

The document raises a practical implementation problem for a mean reversion strategy on multiple cointegrated assets. The cited model estimates an optimal inventory from regression estimates and their derivatives, but its output consists of fractional asset positions. Futures trading instead requires positions in whole contracts, so the author asks how to convert the theoretical inventory into executable quantities for high frequency trading.

The post notes that simply scaling all target positions to make them whole-number quantities would raise trading cost in the author's setting. It does not provide a rounding algorithm, cost model, or tested execution policy, so it cannot establish a best practice. Any conversion would need to account for contract units and the relative exposures implied by the cointegration model; naive rounding could alter hedge relationships. The material is therefore a useful statement of the gap between continuous model outputs and discrete futures orders, rather than a solved strategy.

Key ideas

• The cited strategy estimates optimal inventory across multiple cointegrated assets.
• Regression estimates and their derivatives inform the model's fractional target positions.
• Futures positions must be expressed in integer contract quantities.
• Scaling positions to avoid fractions may increase trading costs.
• The post asks about rounding but supplies no method or empirical evidence.

Tags

# Assets cointegration: optimal inventory for futures market


# Assets cointegration: optimal inventory for futures market












I'm interested in HFT cointegration stratagies, and recently found interesting article "Algorithmic Trading of Co-Integrated Assets"

The article describes mean-reversion strategy for N-assets. Authours build the model which suggests optimal inventory at each moment in time based on regression estimates and their derivatives.

I reproduced the model in matlab and it suggests fractions of assets as optimal inventory. However I am trading on futures market, and it is only allowed to trade with integer amount of contracts. What is the best practice to round fractional number of futures for real HFT trading (simply multiply by 10 is too expensive)?

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.