Analysing the instability in the multivariate correlation structure, the present thesis starts from assessing in-sample and out-of-sample performances of multivariate GARCH models with or without a structural break. The result emphasizes the importance of correlation change point detection for model fittings. We then propose semi–parametric CUSUM tests to detect a change point in the covariance structures of non–linear multivariate models with dynamically evolving volatilities and correlations. The asymptotic distributions of the proposed statistics are derived under mild conditions. Our simulations show that, even though the nearly unit root property distorts the size and power of tests, the standardization of the data with conditional standard deviations in multivariate volatility models can correct such distortions. Lastly, concerning classical trimmed issue in change point test, we extend the semi-parametric CUSUM to weighted CUSUM tests, which enhances the power across either ends of a sample. A Monte Carlo simulation study suggests that weighted CUSUM tests exhibit better performances than unweighted ones in finite samples. Regarding empirical applications, we show the absorption ratio is a leading indicator of the financial fragility, and we study global financial contagion effect, also we investigate unexpected events in the U.S. equity market.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:723392 |
| Date | January 2017 |
| Creators | Zhao, Yuqian |
| Publisher | University of Birmingham |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://etheses.bham.ac.uk//id/eprint/7739/ |
Page generated in 0.002 seconds