Executing large orders in a microscopic market model

TL;DR

This paper extends a market impact model for large order execution by incorporating order size-dependent recovery speeds, improving the realism of optimal trading strategies in microscopic market models.

Contribution

It introduces the GAFS model, a generalization of the AFS model, accounting for variable market recovery speeds based on order size, and derives corresponding optimal strategies.

Findings

The original AFS model's assumption of constant recovery speed is inconsistent with agent-based simulations.

The GAFS model accurately incorporates size-dependent recovery speeds.

Even the improved GAFS model's strategies do not fully match real market behavior.

Abstract

In a recent paper, Alfonsi, Fruth and Schied (AFS) propose a simple order book based model for the impact of large orders on stock prices. They use this model to derive optimal strategies for the execution of large orders. We apply these strategies to an agent-based stochastic order book model that was recently proposed by Bovier, \v{C}ern\'{y} and Hryniv, but already the calibration fails. In particular, from our simulations the recovery speed of the market after a large order is clearly dependent on the order size, whereas the AFS model assumes a constant speed. For this reason, we propose a generalization of the AFS model, the GAFS model, that incorporates this dependency, and prove the optimal investment strategies. As a corollary, we find that we can derive the ``correct'' constant resilience speed for the AFS model from the GAFS model such that the optimal strategies of the AFS and the GAFS model coincide. Finally, we show that the costs of applying the optimal strategies of the GAFS model to the artificial market environment still differ significantly from the model predictions, indicating that even the improved model does not capture all of the relevant details of a real market.