A multi time-scale statistical feedback model of volatility: Stylized facts and implications for option pricing
<-- Return to the list
Date: 10-10-2005
Start Time:
6:00pm
End Time: 7:30pm
Speaker: Lisa Borland, EVA Funds, San Francisco
Location: 412 Schapiro CEPSR, Davis Auditorium
ABSTRACT
We present a model of volatility in financial markets which, in a simple and intuitive way, captures anomalous and subtle statistics and dynamics of real price time series. We find that by incorporating statistical feedback across a continuum of time-scales one can reproduce and predict many empirical stylized facts with a small set of parameters. Furthermore, these parameters can be calibrated to data in a seemingly "universal" fashion. Option pricing based on this model - which can also be seen as a kind of long-memory ARCH process - is discussed. At one limit the non-Gaussian option price model (which was the subject of some of our previous work) is recovered, for which closed form solutions are obtained. Applications of the latter model to the pricing of stock options and credit default swaps are briefly discussed.
BIO
Lisa Borland holds a Ph.D. in theoretical physics from the University of Stuttgart, Germany. She has been working in finance since 1999, and is currently at Evnine-Vaughan Associates Inc., a San-Francisco based hedge fund. Her main focus lies in the research and development of new trading strategies.