Abstract / Description of output
A novel multivariate factor GARCH specification is used to obtain conditional covariance matrices of minimum variance portfolios containing a very large number of assets. The approach allows for time varying factor loads, and achieves great flexibility by allowing alternative specifications for the covariance among factors and for the variance of the asset-specific part of return. Minimum variance portfolios based on the proposed conditional covariance matrix specification are shown to deliver less risky portfolios in comparison to benchmark models, including existing factor approaches.
Original language | English |
---|---|
Pages (from-to) | 606-617 |
Number of pages | 12 |
Journal | Computational Statistics and Data Analysis |
Volume | 76 |
DOIs | |
Publication status | Published - 1 Aug 2014 |
Keywords / Materials (for Non-textual outputs)
- dynamic conditional correlation (DCC)
- forecasting
- Kalman filter
- learning CAPM
- performance evaluation
- sharpe ratio