Abstract
To evaluate some Value at Risk models, the empirical likelihood approach to martingales is recommended. It turns out that the usual Wilks' theorem still holds in this case under mild conditions, and then it can be performed easily. Simulations were carried out for examining the performance of the new method.
| Original language | English |
|---|---|
| Pages (from-to) | 455-468 |
| Number of pages | 14 |
| Journal | Statistica Sinica |
| Volume | 20 |
| Issue number | 1 |
| Publication status | Published - Jan 2010 |
Scopus Subject Areas
- Statistics and Probability
- Statistics, Probability and Uncertainty
User-Defined Keywords
- Empirical likelihood
- Martingale
- Non-nested test
- Specification test
- Value at Risk