Copula structural shift identification

NRU Higher School of Economics , 2012. No. 05.

Бродский Б. Е., Пеникас Г. И., Сафарьян И.

This paper aims at presenting the research results of revealing a structural shift in copula-models of multivariate time-series. A nonparametric method of structural shift identification and estimation is used. The asymptotical characteristics (the probabilities of the I-type and II-type errors, and the probability of the estimation error) of the proposed method are analyzed. The simulation method verification results for Clayton and Gumbel copulas are presented and discussed. The empirical part of the paper is devoted to structural shift identification for multivariate time series of interest rates for Euro-, US Dollar- and Ruble-zones. The empirical application provides strong evidence of the efficiency for the proposed method of structural shift identification.

Ключевые слова: copula, Kolmogorov-Smirnov statistics, interest rates, structural shift, финансовая наука, денежные и налоговые теории, кредитно-финансовые институты