Impact of systemic risk measures on portfolio diversification: Evidence from the Johannesburg Stock Exchange

September 14, 2023 · Christian
Impact of systemic risk measures on portfolio diversification: Evidence from the Johannesburg Stock Exchange

Anaclet Kipupi Kitenge and J.W. Muteba Mwamba

SUMMARY

This paper develops a domestic portfolio diversification strategy aimed at optimising investment capital during market downturns. The strategy consists in constructing two sub-portfolios (the adverse returns portfolio referred to here as ARP, and the favourable returns portfolio referred to here as FRP) using a combination of copulas, extreme value theory (EVT) and the GARCH-based conditional value-at-risk (CVaR). Under certain specific assumptions, a quadratic mean-variance optimisation framework is implemented to obtain the optimal weights. Using the daily returns of nine sector indices of the Johannesburg Stock Exchange, the paper concludes, firstly, that the sectors most correlated with the stock market (proxied by the ALSI) are the ones that contribute most to maximising the ARP, while the sectors offering lower returns with higher risk are the ones that contribute most to maximising the FRP. Secondly, the paper finds that the efficient portfolio outperforms the benchmark portfolio when financial markets are in crisis; however, the converse is true when financial markets are in a recovery phase.

Keywords: Portfolio diversification, Systemic risk, Quadratic mean-variance portfolio, CVaR, Spearman correlation coefficient, CARA
JEL codes: C46, C53, C58, C61, C65, G01, G11


ABSTRACT

This paper develops a domestic portfolio diversification strategy that optimizes investment capital during market downturns. The strategy consists in constructing two sub-portfolios (the adverse returns portfolio herein referred to as ARP, and the favorable returns portfolio herein referred to as FRP) using a combination of copulas, extreme value theory distribution, and the GARCH1 -based conditional value-at-risk (CVaR). Under some specific assumptions, a quadratic mean-variance optimization framework is implemented to obtain the optimal weights. Using daily returns of nine Johannesburg Stock Exchange sector indices, the paper finds firstly that sectors that are more correlated to the stock market (proxied by the ALSI) are the ones that contribute more in maximizing the ARP, and sectors that have lower returns with higher risk are the ones that contribute more in maximizing the FPR. Secondly, the paper finds that the efficient portfolio has a better performance than the benchmark portfolio when the financial markets are in turmoil; however, the converse is true when the financial markets are in upturns.

Keywords: Portfolio diversification, Systemic risk, Quadratic mean-variance portfolio, CVaR, Spearman correlation coefficient, CARA
JEL codes: C46, C53, C58, C61, C65, G01, G11