A warning tool for financial time-series based on time dependent Generalized Hurst exponents

The study of scaling in financial systems has been a field of investigation for many years now. Previous studies have shown that financial series, especially from stock-markets, display multiscaling, which is nowadays widely accepted as an empirical stylized fact of financial time series. However, scaling in a financial time series has also been shown to vary with time. For example, there have been studies trying to link this variation with dynamical elements in the underlying title such as, for example, with the development of stock-market bubbles, and predictability of an index raising the question whether scaling analysis can be used as a signaling tool for financial markets. In this work, we attempt to contribute towards this question by applying a visual methodology to algorithmically detect critical changes in the scaling of the underlying complex time-series based on the Generalized Hurst Exponent (GHE) method. The methodology involves the degree of multiscaling at a particular time instance, the multiscaling trend evaluated by the Change Point Analysis method and a rigorous evaluation of the statistical significance of the results. Applying this methodology to a number of stock market indices we have identified particular patterns in the temporal co-evolution of the different GHE time-series, H(q), for various values of the parameter q (figure 1). These GHE patterns, distinguish in a statistically robust way, not only between time periods of uniscaling and multiscaling, but also among different types of multiscaling: symmetric multiscaling (M) and asymmetric multiscaling (A). Transitions from uniscaling to multiscaling behavior occur before critical market events, such as stock market bubbles. Moreover, particular asymmetric multiscaling patterns appear during critical stock market eras and provide useful information about market conditions suggesting that the method could be used as a signaling tool for future critical market events.

Συνεδρία: 
Authors: 
Ioannis Antoniades, Giuseppe Brandi, Lykourgos Magafas and Tiziana Di Matteo
Room: 
2
Date: 
Thursday, December 10, 2020 - 18:00 to 18:15

Partners

Twitter

Facebook

Contact

For information please contact :
ccs2020conf@gmail.com