Portfolio optimization from a copulas-GJR-GARCH-EVT-CVAR model: empirical evidence from ASEAN stock indexes

Abstract

This study employs several methods to simulate and construct the portfolio from stock indexes of the six Association of Southeast Asian Nations (ASEAN) markets during the period from January 2001 to December 2017, namely, time-varying Copulas; Glosten, Jagannathan and Runkle (GJR); generalised autoregressive conditional heteroskedasticity (GARCH); extreme value theory (EVT); and conditional value at risk (CVaR). Our target is minimising the risk based on CVaR, then achieving the maximal expected return for investors. Our model also sheds further light on the role of the dependence structure among stock indexes by employing elliptical (student t) Copulas, which are incorporated for simulating the optimal portfolios. Our findings suggest that the investor should invest in the optimal portfolio, which lies in the efficiency curve. Hence, the optimal portfolio has similar time-varying characteristics across the dependence of Copulas, as well as confidence levels. The research implications can be employed practically by portfolio managers and individual investors who desire to invest in ASEAN equity markets. Therefore, our findings can draw investors' attention to constructing the portfolio with the dependence level via time-varying Copulas and minimise the risk represented by CVaR rather than traditional variance.