Optimization of Financial Asset Portfolio Using GARCH-EVT-Copula-CVaR Model

Abstract

Since the pioneering work of Markowitz on portfolio theory in 1950s, numerous developments have advanced to improve the original technique of portfolio optimisation. Current research on the topic focuses on integrating models to capture real-world financial characteristics, including volatility clustering, heavy tails, and non-linear dependencies. Portfolio optimisation is a critical aspect of financial risk management, requiring sophisticated models to accurately assess risk and optimise asset allocation. This study implemented an integrated approach utilising Generalised Autoregressive Conditional Heteroskedasticity (GARCH) for volatility estimation, Extreme Value Theory (EVT) for modelling extreme market movements, Copula functions for capturing dependencies between financial assets, and Conditional Value at Risk (CVaR) for robust risk assessment in the portfolio of financial assets. By applying this methodology to a portfolio of financial assets, the empirical results demonstrate that the GARCH-EVT-Copula-CVaR model significantly improves risk estimation, portfolio selection, and optimisation. The empirical results also confirm its superiority over conventional models, highlighting its potential for enhanced risk management and portfolio asset allocation. The integrated model can be recommended for utilisation by stakeholders in the financial markets, investors, and regulators for policy formulation and informed decision-making.