Proposing a portfolio optimization model based on the GARCH-EVT-Copula combined approach

Document Type : Research Paper


1 Department of Financial Management, Central Tehran Branch, Islamic Azad University, Tehran, Iran.

2 Department of Industrial Management, Central Tehran Branch, Islamic Azad University, Tehran, Iran

3 Department of Financial Management, Central Tehran Branch, Islamic Azad University, Tehran, Iran


This study aims at optimizing the portfolio of financial assets and in particular focuses on the stock market with conditional value at risk (CVaR) as the portfolio risk measure. This study uses generalized conditional heterogeneity variance methods, the dependency structure, the extreme value theory, and with the GARCH-EVT-Vine-Copula approach to optimize the portfolio and minimize the CVaR of a stock portfolio during a certain period by the re-weighting method. Modeling is based on the performance data of 7 companies among the top 50 listed companies during the period 2015 to 2021. The results show that considering the extreme values and structural dependence between the examined time series improves the risk identification between these markets. In addition, among the studied models, the out-of-sample results for the accumulated wealth function of different models show that when considering the dependence structure, the EGARCH-EVT model based on the Coppola Vine function results outperforms other models.


[1] K. Aas, C. Czado, A. Frigessi and H. Bakken, Pair-copula constructions of multiple dependence, Insurance: Math. Econ. 44 (2009), no. 2, 182–198.
[2] H. Boubaker and N. Sghaier, Portfolio optimization in the presence of dependent financial returns with long memory: A copula based approach, J. Bank. Finance 37 (2013), no. 2, 361–377.
[3] P. Christoffersen, P. Christoffersen, V. Errunza, K. Jacobs, K. Jacobs and X. Jin, Correlation dynamics and international diversification benefits, Int. J. Forecast. 30 (2014), no. 3, 807–824.
[4] D.G. Clayton, A model for association in bivariate life tables and its application in epidemiological studies of familial tendency in chronic disease incidence, Biometrika 65 (1978), no. 1, 141–151.
[5] J. Dismann, E.C. Brechmann, C. Czado and D. Kurowicka, Selecting and estimating regular vine copulae and application to financial returns, Comput. Statist. Data Anal. 59 (2013), 52–69.
[6] P. Embrechts, F. Lindskog and A. McNeil, Modelling dependence with copulas, Rapport Technique, Departement de math´ematiques, Institut Federal de Technologie de Zurich, Zurich, 14 (2001), 1–50.
[7] J. Estrada, Mean-semivariance behavior: Downside risk and capital asset pricing, Int. Rev. Econ. Financ. 16 (2007), no. 2, 169–185.
[8] M.F. Fallah and A. Alizadeh, Investigating performance of the extreme value theory and GARCH-Copula models in predicting the value at risk and expected fall of the portfolio in investment companies of Tehran Stock Exchange, Quarterly J. Financ. Eng. Secur. Manag. 12 (2021), no. 46, 340–364.
[9] S. Fallahpour and E. Ahmadi, Estimating the value at risk of oil and gold portfolio using GARCH-Copula models, Financ. Res. 16 (2014), no. 2, 309–326.
[10] S. Fallahpour and M. Baghban, Application of the CVaR-Capula in stock portfolio optimization and its comparative study with Mean-CVaR method, Quart. J. Econ. Res. Policy 22 (2014), no. 72, 155–172.
[11] M.J. Frank, On the simultaneous associativity of F(x, y) and x+y−F(x, y), Aeq. Math. 19 (1979), no. 1, 194–226.
[12] M. Ghandehari, A. Shamshiri and S. Fathi, Stock portfolio optimization based on non-parametric estimation methods, Prod. Oper. Manag. Quart. 8 (2017), no. 1, 175–184.
[13] E.J. Gumbel, Bivariate exponential distributions, J. Amer. Statist. Assoc. 55 (1960), no. 292, 698–707.
[14] Y. Han, P. Li and Y. Xia, Dynamic robust portfolio selection with copulas, Finance Res. Lett. 21 (2017), 190–200.
[15] M. Karmakar, Dependence structure and portfolio risk in Indian foreign exchange market: A GARCH-EVTCopula approach, Quart. Rev. Econ. Finance 64 (2017), 275–291.
[16] G. Keshavarz Haddad and M. Heirani, Estimating value at risk with dependence structure between financial returns: A copula function-based approach, J. Econ. Res. 49 (2014), no. 4, 869–902.
[17] A. Krzemienowski and S. Szymczyk, Portfolio optimization with a copula-based extension of conditional value-atrisk, Annal. Operat. Res. 237 (2016), no. 1-2, 219–236.
[18] C. Luo, L. Seco and L.L.B. Wu, Portfolio optimization in hedge funds by OGARCH and Markov switching model, Omega 57 (2015), 34–39.
[19] H. Markowitz, The utility of wealth, J. Politic. Econ. 60 (1952), no. 2, 151–158.
[20] D.B. Nelson, Conditional heteroskedasticity in asset returns: A new approach, Econometrica: J. Econ. Soc. 59 (1991), no. 2, 347–370.
[21] A. Patton, On the out-of-sample importance of skewness and asymmetric dependence for asset allocation, J. Financ. Econ. 2 (2004), no. 1, 130–168.
[22] E. Pishbahar and S. Abedi, Calculating the portfolio value at risk: Capillary approach, Quarterly J. Financ. Eng. Secur. Manag. 8 (2017), no. 30, 55–73.
[23] S.-H. Poon, M. Rockinger and J. Tawn, Modelling extreme-value dependence in international stock markets, Statist. Sinica 13 (2004), 929–953.
[24] A. Pouyanfar and S.H. Mousavi, Estimating the value at risk of overnight data with EVT-Copula approach, Quart. J. Risk Model. Financ. Engin. 1 (2015), no. 2, 129–144.
[25] M. Sahamkhadam, A. Stephan and R. Ostermark, Portfolio optimization based on GARCH-EVT-Copula forecasting models, Int. J. Forecast. 34 (2018), no. 3, 497–506.
[26] A. Sklar, Fonctions de repartition an dimensions et leurs marges, Publ. Inst. Statist. Univ. Paris 8 (1959), 229–231.
[27] P. Xue-Kun Song, Multivariate dispersion models generated from Gaussian copula, Scand. J. Statist. 27 (2000), no. 2, 305–320.
Volume 14, Issue 6
June 2023
Pages 197-210
  • Receive Date: 13 April 2022
  • Revise Date: 07 June 2022
  • Accept Date: 28 June 2022