Asymmetric downside risk across different sectors of the US equity market

Abstract

This study integrates a threshold-mean equation with an asymmetric power autoregressive conditionally heteroscedastic (APARCH) model to examine the behavior of sector-specific exchange-traded funds (ETFs) during extreme market downturns between December 23, 1998, and November 2, 2022. Thus, predetermined and optimal boundary points are applied to the extreme left tail of the return distribution to assess the extent of downside risk differentials inside and outside the extreme drawdown zone without splitting the sample period. According to the findings, the betas of the ETFs XLI, XLP, XLV, and XLY are comparable under both extreme and nonextreme market conditions. In contrast, XLF, XLE, and XLU have higher downside betas during extreme market conditions compared with their nonextreme betas, while XLB and XLK exhibit the opposite pattern. The results remain robust and consistent regardless of how the boundary points are established. The estimated models in this study were successfully subjected to a series of diagnostic tests, suggesting that the commonly held view about asymmetric responses in different market conditions does not apply to all market segments.