Speaker:  Dr Maggie Chen, Cardiff University
Chair: Professor Victor Murinde, SOAS University of London
Time: Wednesday, 29 January 2020, 13:00-15:00
Venue: S314 (Paul Webley Wing, Senate House), SOAS University of London

Abstract

Jump detection in financial time series has evolved over the past few decades due to several factors. The first is increased empirical evidence of the impact of jumps on returns and volatility. The second reason is the increased frequency of jumps with the advent of high-frequency finance and computer-to-computer trading. Third, and as a result, risk assessment, stop-loss positions, portfolio rebalancing, flash crashes and other ramifications may be triggered by jumps. Jump detection methods, however, have often treated a jump as a singular, random and isolated shock significantly standing out from the rest of the values of the time series. Most jump detection techniques were not designed to capture consecutive jumps, which may constitute a massive shock to the financial system even if the individual jumps are relatively modest in size. More importantly, successive jumps often reflect a behavioral contagion among market agents, be they human or machines. Thus, from the financial-economic perspective we advance the BCH method to account more accurately for both singular jumps and consecutive jumps. This judges the size of individual returns with a measure of local volatility based on running medians of absolute returns.  We also compare this method with existing methods such as the RV-BV method, or Lee and Mykland (2008). The comparison is carried out on empirical S\&P 500 data as well as simulated data. We find that the BCH method generally out-performs other jump detection methods.

Seminar Presenter

Dr Jing (Maggie) Chen is a Reader in Financial Mathematics in Cardiff University. She also served in Swansea University, Columbia University (NYU) and University College London. Her research interests are on market microstructure, Volatility & financial jump, complex financial systems and stability, utilizing theory-underpinned modelling techniques to provide robust empirical evidence. She is also one of the pioneers to introduce Hawkes processes into finance applications, which has led her to edit a couple of special issues for major finance journals recently.