Résumé
•The paper investigates volatility dynamics in U.S. and European equity markets using the ARFIMA-HYAPGARCH-M framework, which integrates dual long memory, asymmetric volatility, and risk premium effects.•The analysis distinguishes between two distinct periods: the pre-Covid-19 period and the Covid-19 period, enabling a focused examination of the pandemic’s shock impact.•The skewed Student-t distribution provides a more robust modeling of fat-tailed and skewed returns data, enabling to handle extreme market shocks.
Financial data exhibit distinctive characteristics known as stylized facts including volatility clustering, long memory, the leverage effect, and risk premium.
In this paper, we introduce a innovative volatility model (ARFIMA-HYAPGARCH-M) designed to effectively capture these features in both the S&P 500 and the European STOXX600 indices, before and during the Covid-19 pandemic.
Empirical findings reveal a significant surge in return volatility across both U.S. and European stock markets during the pandemic. Moreover, the data exhibit dual long memory properties in both the mean and variance of returns, along with an evidence of asymmetry and the leverage effect. Furthermore, the results show that risk premiums increased during the Covid period, confirming that investors demand higher compensation during periods of “bad” volatility compared to periods of “good” volatility.
As such, the ARFIMA-HYAPGARCH-M volatility model provides a valuable tool for improved risk assessment, enabling investors and portfolio managers to make more informed decisions. Additionally, the model can enhance the performance of hedging strategies by accurately capturing volatility dynamics.