Abstract
A cointegrated vector AR-GARCH time series model is introduced. Least squares estimator, full rank maximum likelihood estimator (MLE), and reduced rank MLE of the model are presented. Monte Carlo experiments are conducted to illustrate the finite sample properties of the estimators. Its applicability is then demonstrated with the modeling of international stock indices and exchange rates. The model leads to reasonable financial interpretations. Copyright © 2005 The Institute of Statistical Mathematics.
| Original language | English |
|---|---|
| Pages (from-to) | 83-103 |
| Journal | Annals of the Institute of Statistical Mathematics |
| Volume | 57 |
| Issue number | 1 |
| DOIs | |
| Publication status | Published - Mar 2005 |
Citation
Wong, H., Li, W. K., & Ling, S. (2005). Joint modeling of cointegration and conditional heteroscedasticity with applications. Annals of the Institute of Statistical Mathematics, 57(1), 83-103. doi: 10.1007/BF02506881Keywords
- Cointegration
- Full rank maximum likelihood estimator
- Least squares estimator
- Partially nonstationary
- Reduced rank MLE
- Vector AR-GARCH model