Portfolio Optimization with PCC-GARCH-CVaR model

Abstract

This thesis investigates the Conditional Value-at-Risk (CVaR) portfolio optimization approach combined with a univariate GARCH model and pair-copula constructions (PCC) to determine the optimal asset allocation for a portfolio. The methodology focuses on minimizing CVaR as the risk measure in replacement of variance used in the traditional optimization framework of Markowitz. GARCH model provides a tool for predicting and analyzing the time-varying volatility financial assets are exposed to, while copulas allow us to model the non-linear dependence structure and margins separately. We compare the performance of the CVaR optimized portfolio with other investment strategies such as Constant-Mix and Buy-and-Hold. Although the selection of strategy depends on the investor risk profile, it is empirically shown that the proposed CVaR optimized portfolio outperforms the other two investment strategies based on the accumulated wealth in the long run.