Copulas for Markovian dependence
2010 (English)In: Bernoulli, ISSN 1350-7265, Vol. 16, no 2, 331-342 p.Article in journal (Refereed) Published
Copulas have been popular to model dependence for multivariate distributions, but have not been used much in modelling temporal dependence of univariate time series. This paper demonstrates some difficulties with using copulas even for Markov processes: some tractable copulas such as mixtures between copulas of complete co- and countermonotonicity and independence (Frechet copulas) are shown to imply quite a restricted type of Markov process and Archimedean copulas are shown to be incompatible with Markov chains. We also investigate Markov chains that are spreadable or, equivalently, conditionally i.i.d.
Place, publisher, year, edition, pages
2010. Vol. 16, no 2, 331-342 p.
copulas, exchangeability, Markov chain, Markov process
IdentifiersURN: urn:nbn:se:su:diva-48983DOI: 10.3150/09-BEJ214ISI: 000278910700002OAI: oai:DiVA.org:su-48983DiVA: diva2:376250
authorCount :12010-12-102010-12-102010-12-14Bibliographically approved