Vine copula-cross entropy evaluation of dependence structure and financial risk in agricultural commodity index returns

Abstract

Many studies used the empirical Kendall's tau to select a preferable ordering of vine copulas or to fix such a sequence. In this study, for high dimension vine copulas, we propose the vine copula based cross entropy method to figure out a more appropriate ordering of the vine copula. The goal of this study is to estimate the non-conditional, conditional, and tail dependences for agricultural price index returns by using the C-vine and D-vine copula based cross entropy model. In addition, we show that a framework uses the Monte Carlo simulation and the results of vine copula to estimate the expected shortfall (ES) of an equally weighted portfolio. The optimal portfolio allocations can also be estimated using global optimization with the differential evolution algorithm. © Springer International Publishing Switzerland 2014.