Nonlinear Shift Contagion Modeling: Further Evidence from High Frequency Stock Data - Université de Rennes Accéder directement au contenu
Chapitre D'ouvrage Année : 2011

Nonlinear Shift Contagion Modeling: Further Evidence from High Frequency Stock Data

Résumé

This paper investigates the contagion hypothesis for the French and German stock markets using a combination of a Switching Transition Error Correction model and a Generalized Autoregressive Conditional Heteroscedasticity (STEC-GARCH) model. The main advantage of this double nonlinear error-correction modeling is to specify a time-varying process that apprehends the dynamic evolution of the contagion and reproduces its speed, its extreme regimes as well as its intermediate states, by taking into account the possible linkages between these markets. More importantly, these techniques capture two kinds of nonlinearity: nonlinearity in the mean and nonlinearity in the variance. Applying this modeling on the intraday data of the CAC40 and DAX100 indices over the pre-crisis period (2004-2006) and the post-crisis period (2007-2009), our results indicate significant shift contagion between studied markets. There is also evidence of nonlinear time-varying error correcting-mechanism toward the long-run equilibrium.
Fichier non déposé

Dates et versions

halshs-00601428 , version 1 (17-06-2011)

Identifiants

  • HAL Id : halshs-00601428 , version 1

Citer

M.H. Arouri, F. Jawadi, Waël Louhichi, D. K. Nguyen. Nonlinear Shift Contagion Modeling: Further Evidence from High Frequency Stock Data. G.N. Gregoriou and R. Pascalau (Eds.). Financial Econometrics Handbook, Palgrav Macmillan London, pp.143-160, 2011. ⟨halshs-00601428⟩
115 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More