A characterisation of cross-impact kernels

TL;DR

This paper characterizes cross-impact kernels in financial markets, focusing on models that ensure no arbitrage and martingale properties, with practical calibration methods demonstrated on SP500 futures data.

Contribution

It introduces a unified framework for kernel-based cross-impact models, identifying conditions for arbitrage-free and martingale-admissible kernels, and provides calibration formulas.

Findings

Identified the intersection of arbitrage-free and martingale-admissible kernels.

Derived formulas for calibrating cross-impact kernels from data.

Validated the models using SP500 futures data.

Abstract

Trading a financial asset pushes its price as well as the prices of other assets, a phenomenon known as cross-impact. We consider a general class of kernel-based cross-impact models and investigate suitable parameterisations for trading purposes. We focus on kernels that guarantee that prices are martingales and anticipate future order flow (martingale-admissible kernels) and those that ensure there is no possible price manipulation (no-statistical-arbitrage-admissible kernels). We determine the overlap between these two classes and provide formulas for calibration of cross-impact kernels on data. We illustrate our results using SP500 futures data.