报告题目：Flash Crashes, Jumps and Running Jumps: A New Method for Jump Detection
主讲人：Prof. Jing Chen, Cardiff University, UK
摘要：While studying the S & P 500 financial index, recorded every two minutes, we apply a well-known method of identifying jumps based on the difference between Realized Volatility and Bi-power Variation in each trading day. Surprisingly, we find no indication that jumps occurred on flash crash day (May 06, 2010) and similar days. The problem is that this method cannot tolerate large variations in consecutive intervals, which can cause the bi-power variation to exceed the realized volatility, and consequently leads to a failure to detect jumps. We, therefore, propose a new jump detection method based on a daily volatility measure that uses a median approach to avoid the masking effects caused by multiple large returns. We further introduce running jumps aimed explicitly at studying the occurrence of sequences of large neighbouring returns over successive intervals that effectively form a single jump. Our method captures significant jumps on flash crash day and other days when market events trigger sharp price changes.
主讲人概况：Dr Jing (Maggie) Chen is, currently, Reader of Financial Mathematics in the School of Mathematics, Cardiff University. She holds BSc in Computer Science from the Lanzhou Jiaotong University, China, MSc in Finance & Investment Management and PhD in Finance from the University of Aberdeen, UK (2011). Maggie holds visiting/adjunct professorship in the Statistics Departments in both the Columbia University (New York City) and University College London.
Maggie’s main research focus is on liquidity provision, information dynamics and their relation to policy changes in financial markets. Her work has appeared in esteemed international scholarly journals including the Pacific Basin Finance Journal, European Journal of Finance, Journal of Economics Behaviour and Organisation and Journal of Forecasting. Maggie has several roles including associate editor of Cogent Economics, guest editors for special issues on Hawkes processes and Finance for both Quantitative Finance and The European Journal of Finance.