Black-Litterman model with copula-based views in mean-CVaR portfolio optimization framework with weight constraints

This study examines the portfolio optimization problem by exploiting daily data of 10 international Exchange Trade Funds (ETF) from 2012 to 2022. We extend Black-Litterman (BL) approach using ARMA-GARCH-copula-based expected returns as a proxy for investor views and use CVaR metric risk measure in procedure. The BL provides Bayesian methodology combining equilibrium produce returns. Regular Vine (R-vine) copula since it flexible multivariate dependency modeling. suggested is compared against copula-CVaR portfolio, which likewise avoids excessive corner solutions that many approaches would generate case extreme values estimated parameters. compare performance these two out-of-sample back-testing benchmarks: Mean–Variance optimizations (MV) equal weights (EW). To further reduce sensitivity considered strategies input parameters, we evaluate at three levels maximum weight constraints: 30%, 40%, 50%. Moreover, this paper, consider different view confidence—τ model significantly affects obtained results inferences. calculate report portfolios’ tail risks, drawdown, turnover, break-even point all approaches. Our empirical analysis indicates better CBL regarding lower higher risk-adjusted returns, turnover point.