Back to articles of current journal
PDF: Author(s): Antonov I. N.,
Number of journal: 1(34) Date: February 2016

Linear dependence often use for forecasting the exchange rate of foreign currencies; however, these methods are often ineffective. In this article the forecast of foreign currencies (US dollar, Euro, pound sterling, Canadian dollar and Australian dollar) is made on the multidimensional Archimedean copula-models. Copulas Clayton, Frank and GumbelHougaard were built and estimated by maximum likelihood and Bayesian method with Metropolis-Hastings. Selection of the model and the estimated parameters were done by Kolmogorov-Smirnov test. As the result the best model has been Clayton copula estimated by maximum likelihood. The best forecast was obtained for Canadian dollar.


Archimedean copulas, Bayesian estimation, currency risk, currency, Clayton copula, Frank copula, Gumbel-Hougaard copula, maximum likelihood, multidimensional copula, forecast

For citation:

Antonov I. N. Forecasting currency rate by copula-models // Business. Education. Law. Bulletin of Volgograd Business Institute. 2016. No. 1 (34). P. 158–164.