COPULA MODELS OF THE JOINT DISTRIBUTION OF EXCHANGE RATES
The paper aims at investigating the joint distribution of currency rates using HAC, HKC and Vine copulas in several time periods. Models were constructed using Archimedean copulas including Gumbel-Hougaard, Joe BB1 and Frank copulas, and their parameters were estimated by maximum likelihood. The bes...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | Russian |
| Published: |
Novosibirsk State University Press
2016-12-01
|
| Series: | Mir Èkonomiki i Upravleniâ |
| Subjects: | |
| Online Access: | https://nsu.ru/ef/vestnik_ngu_ef/2016_4_2 |
| Summary: | The paper aims at investigating the joint distribution of currency rates using HAC, HKC and Vine copulas in several time periods. Models were constructed using Archimedean copulas including Gumbel-Hougaard, Joe BB1 and Frank copulas, and their parameters were estimated by maximum likelihood. The best models were built using hierarchical Archimedean copulas and the worst were obtained with vine copulas. In comparison with the HAC, the main advantage of hierarchical Kendall copulas is the possibility to use a two-parameter copula Joe BB1. The best models were obtained with Frank copula, while Gumbel-Hougaard copula has shown a decent result only in the third period. Additionally, in this paper was made an attempt to get the forecast of exchange rates using Kendall’s and Marshall-Olkin’s algorithms. The most accurate forecast was obtained with Gumbel-Hougaard copula for euro and franc. |
|---|---|
| ISSN: | 2542-0429 2542-0429 |