Portfolio optimization based on forecasting models using vine copulas: An empirical assessment for global financial crises

We employ and examine vine copulas in modeling symmetric and asymmetric dependency structures and forecasting financial returns from 2001 to 2022, a period that includes the 2008 financial crisis, the 2011 European sovereign debt crisis, the 2020 COVID-19 pandemic crisis, and the 2022 Russian invasion of Ukraine with the resulting energy crisis. We analyze the asset allocations performed and test different portfolio strategies, such as maximum Sharpe ratio, minimum variance, and minimum conditional value at risk. Using international financial market indices, we specify the regular, drawable, and canonical vine copulas, such as the Gaussian, Student's t$$ t $$, Clayton, Frank, Joe, Gumbel, and mixed copulas, and analyze both in-sample and out-of-sample portfolio performances. Out-of-sample portfolio back-testing shows that vine copulas reduce portfolio risk better than the benchmark portfolio strategies and also better than simple multivariate copulas. Overall, we find that mixed vine copula models perform best with regard to risk reduction, both for the entire period 2001-2022 and during financial crises periods. Thus, a mixture of symmetric and asymmetric copula families works best in terms of portfolio risk reduction irrespective of the chosen optimization approach.