| Creator |
|
| Language |
|
| Publisher |
|
|
|
| Date |
|
| Source Title |
|
| Vol |
|
| First Page |
|
| Last Page |
|
| Publication Type |
|
| Access Rights |
|
| JaLC DOI |
|
| Abstract |
This paper compares generalised autoregressive conditional heteroscedastic (GARCH) and stochastic variance (SV) modelling approaches in analysing the dynamics of Japan's stock market volatility using ...monthly time series from January, 1984 through April, 2013. GARCH models essentially model the conditional variance of returns given past returns and observations. While SV models, have different data-generating process, with variance that is specified to follow some unobserved stochastic process. We examine estimates of GARCH models with and without breaks accounting for market crash and financial crises to assess their impact on stock market returns volatility and SV-type models. Using a variety of return transformations, we find persistence and variability in the relevant parameters of both GARCH and SV models. We observe that the volatility persistence parameter in the SV model which indicates volatility clustering is comparable with the persistence measure of GARCH models and a similarity in the trend of the estimated SV model with that of the IGARCH model. Finally, we investigate whether GARCH and SV-type models differ significantly in their ability to predict the volatility of Japan's stock index returns over horizons ranging from 1, 3, 6 to 12 months.show more
|
| Table of Contents |
1. Introduction 2. Previous Studies, Asset Returns Characteristics and Stylised Facts 3. Methodology and Theoretical Framework 4. The Data and Estimation Results 5. Estimation Results and Discussions 6. Volatility Forecasting Methodologies and Prediction 7. Summary and Conclusion
|
Export Mendeley
p001