Pair copula constructions in portfolio optimization ploblem

The choice and estimation of joint probability distribution function are key steps in portfolio optimization problem. As such distribution functions pair-copula constructions (PCC), or vine-copulae, on arbitrary R-vines are used. For the investor with exponential utility criterion the NYSE oil and gas sector-based portfolios are formed. It is shown, that PCC portfolios gain more profit and also PCCs provide reliable VaR estimates. However, on Russian oil and gas stock market PCC portfolio performance is the weakest among competing portfolios. This could be due to shortcomings of maximal spanning trees procedure, which is commonly used to obtain optimal vine structure.