Parts | Days | Selection | Search | Downloads | Help
AKSOE: Physik sozio-ökonomischer Systeme
AKSOE 18: Finanzm
ärkte und Risikomanagement II
AKSOE 18.5: Talk
Thursday, March 14, 2002, 11:30–12:00, H8
Influence of changing volatility on correlations between stock price changes — •Christof Reese and Bernd Rosenow — Institut für Theoretische Physik, Universität zu Köln, Zülpicher Str. 77, 50937 Köln, Germany
Financial time series are characerized by non-stationary standard deviation (volatility). This effect is described by GARCH models relating future volatility to both present volatility and return fluctuations. Correlations between different time-series are affected by changing volatility and can be modeled by multivariate GARCH [1] processes. This requires the estimation of O(N2) parameters where N denotes the number of time-series considered simultanously. From the study of empirical correlation matrices, it is known that a large percentage of correlation coefficients does not contain information [2]. Motivated by this findings we investigate multivariate GARCH processes using methods of Random Matrix Theory.
[1] Engle, R., K. Kroner, Econometric Theory, 11, 122(1995)
[2] Plerou, V., P. Gopikrishnan, B. Rosenow, L.A.N. Amaral, and H.E. Stanley, Phys. Rev. Lett. 83, 1471(1999)