Don't Stop Me Now! Identification and Prediction of Unnecessary Volatility Interruptions
45 Pages Posted: 14 Dec 2024
Abstract
Exchange operators use circuit breakers like volatility interruptions to prevent transitory or error-induced price shocks. However, these mechanisms can impede market efficiency if triggered by legitimate price changes due to new information. We introduce a clustering approach to identify unnecessary volatility interruptions that are triggered within persistent price trends, thereby delaying price discovery. Our findings indicate that such interruptions are more likely to occur when liquidity and order book activity are high and relevant news is present. To improve market design, we propose a deep learning model that can predict unnecessary interruptions based on pre-interruption market and news data.
Keywords: Circuit Breakers, Volatility Interruptions, market quality, price discovery, Market Design, Machine learning
Suggested Citation: Suggested Citation