Title: Portfolio optimization based on forecasting models using vine copulas: An empirical assessment for the financial crisis
Authors: Andreas Stephan - Linnaeus University (Sweden) [presenting]
Maziar Sahamkhadam - Linnaeus University (Sweden)
Abstract: Vine copulas are employed and examined in modeling the symmetric and asymmetric dependency structure and forecasting of financial returns. Asset allocation is performed during the 2007-2010 financial crisis and different portfolio strategies are tested including maximum reward-to-risk ratio, minimum variance and minimum conditional Value-at-Risk. Regular, drawable and canonical vine copulas are specified including Clayton, Frank, Joe and mixed copula. Both in-sample and out-of-sample analyses of portfolio performances are conducted. The out-of-sample portfolio back-testing shows that vine copulas reduce portfolio risk more than simple copulas. Considering portfolio out-of-sample CVaR, Frank and mixed vine copulas result in lower downside risk. The results of the VaR back-testing shows improvement in forecasting of the downside risk for all portfolio strategies obtained from using simple Clayton and mixed copula families, implying time-varying tail dependence of stock market returns. Copula families which capture no tail dependence (Frank) and upper tail dependence (Joe) lead to higher terminal values of portfolios over the financial crisis.