We study the statistical properties of events affecting a limit order book -market orders, limit orders and cancellations. These events are found to exhibit clustering in time as well as significant cross-correlation. We show that these dependencies may be adequately represented by modeling order book events as a multi-dimensional self-exciting point process, for which a tractable parameterization is proposed. Using high-frequency data from the Trades and Quotes database, we perform a Maximum Likelihood Estimation of the model and assess its predictive performance for a variety of stocks.
Speaker: Ekaterina Vinkovskaya, Columbia University
Slides: (TBA)
Event Details
A Point Process Model for the High-Frequency Dynamics of a Limit Order Book
- Event Date: April 26, 2011
- Event End Date: April 26, 2011
- Event Start Time: 10:15 AM
- Event End Time: 11:15 AM
- Event Location: Hill 705
- Event Type: Mathematical Finance and Probability Seminars
- Extra Info: Speaker: Ekaterina Vinkovskaya, Columbia University