Title: Modeling financial durations with limit order book information
Authors: Tomoki Toyabe - Keio University (Japan) [presenting]
Kentaro Asaba - Keio University (Japan)
Teruo Nakatsuma - Keio University (Japan)
Abstract: It is a stylized fact that durations between executions in financial markets have intraday seasonality and autocorrelation. The Autoregressive Conditional Duration (ACD) model has been widely used to capture these characteristics. However, durations are also supposed to be affected by liquidity in the market. We propose a new ACD model that utilizes the limit order book information for reflecting the liquidity. In our empirical analysis, we applied the proposed ACD model to high-frequency stock price data in the Tokyo Stock Exchange and estimated it with an efficient Markov chain Monte Carlo method. We also conducted model comparison among different specifications of the proposed model.