BURSTING BEHAVIOR OF THE NON-LINEAR STOCHASTIC MODELS APPLICABLE TO THE FINANCIAL MARKETS

Aleksejus Kononovicius1, Vygintas Gontis1

1 Institute of Theoretical Physics and Astronomy, Vilnius University, Lithuania

Most of the currently prevailing economic theories were developed rather long time ago. For example classical economic research is considered to be started by A. Smith back in the XVIII century, while the foundations of neoclassical economics were formulated by many prominent researchers in the end of XIX century and in the begging of XX century. Consequently it should not be surprising that these theories are often overly simplistic and tend to disagree with the observations [1, 2]. The situation is very different today - one can obtain loads of the high-frequency empirical data and process it in anyway desired, one can also create complex computer models and build a bridge between theory, experiment and empirical observations.

The empirical analysis enabled mathematicians, physicists and economists to extend the applications of stochastic calculus to the financial markets and create stochastic models [3, 4, 5, 6]. The models have seen various applications and thus they are able to reproduce different aspects of the empirical time series.

Our group has proposed a general class of stochastic differential equations [7],

$$dx = \left(\eta - \frac{\lambda}{2}\right) x^{2\eta - 1} dt + x^{\eta} dW. \tag{1}$$

This class of stochastic differential equations can be used as a base model for the dynamics of the trading activity [8] and absolute returns [9]. The obtained time series successfully replicated the empirical probability and spectral densities. Furthermore this class of stochastic differential equations can be used as a model for the flicker noise observed in very distinct fields.

Our recent work [10] has suggested that the model can be analyzed, and applied in risk management, in the terms of bursting behavior. In this contribution we will discuss the statistics of the above-the-threshold (i.e. burst duration), below-the-threshold (i.e. inter-burst) times and of the other related features of the empirical and model time series. The developed framework might provide insights to the nature of the systemic risk, the fast-growing bubbles and the flash crashes, consequently serving as an excellent tool in the real time policy making.


[1] R. H. Nelson, Economics as Religion: From Samuelson to Chicago and Beyond (University Park: Pennsylvania State University Press, 2001).

[2] J. P. Bouchaud, Economics needs a scientific revolution, Nature 455, 1181 (2008).

[3] K. C. Chan, G. Andrew Karolyi, F. A. Longstaff, A. B. Sanders, An Empirical Comparison of Alternative Models of the Short-Term Interest Rate, The Journal of Finance XLVII, 1209-1227 (1992).

[4] J. A. Hernandez, R. M. Benito, J. C. Losada, An adaptive stochastic model for financial markets, Chaos, Solitons and Fractals 45, 899-908 (2012).

[5] M. Jeanblanc, M. Yor, M. Chesney, Mathematical Methods for Financial Markets (Springer, Berlin, 2009).

[6] H. Zi-Gang, C. Yong, Z. Yong, W. Ying-Hai, Description of dynamics of stock prices by a Langevin approach, Chinese Physics 16, 975 (2007).

[7] J. Ruseckas, B. Kaulakys, Tsallis distributions and 1/f noise from nonlinear stochastic differential equations, Physical Review E 84, 051125 (2011).

[8] V. Gontis, B. Kaulakys, J. Ruseckas, Trading activity as driven Poisson process: comparison with empirical data, Physica A 387, 3891-3896 (2008).

[9] V. Gontis, J. Ruseckas, A. Kononovicius, A long-range memory stochastic model of the return in financial markets, Physica A 389, 100-106 (2010).

[10] V. Gontis, A. Kononovicius, S. Reimann, The class of nonlinear stochastic models as a background for the bursty behavior in financial markets, Advances in Complex Systems 15, 1250071 (2012).