Order Clustering and Liquidity Collapse in Algorithmic Trading Environments

Abstract

Algorithmic trading systems increasingly rely on automated decision engines to execute large volumes of financial transactions. While such systems improve operational efficiency, they also introduce new forms of systemic risk caused by feedback loops and synchronized behavior. This study analyzes transaction logs collected from a quantitative trading platform between 2019 and 2022, covering approximately 3.4 billion orders across equity and futures markets. Abnormal volatility events were examined in relation to order clustering, latency fluctuations, and strategy interactions. Results indicate that under high market stress, order cancellation rates increased by more than 270%, while short-term liquidity dropped by nearly 35%. A simulation model was constructed to reproduce these dynamics under controlled conditions. Findings suggest that minor parameter changes in automated strategies can amplify market instability when combined with network latency and information asymmetry. These results highlight the necessity of system-level risk supervision in large-scale automated trading environments.