CMStatistics 2022: Start Registration
View Submission - CFE
Title: Asymmetric intra-day volatility pattern and price jump detection: Evidence from international equity indices Authors:  Ping Chen Tsai - National Sun Yat-sen University (Taiwan) [presenting]
Abstract: Current methods of price jump detection using high-frequency data typically assume a constant intra-day volatility pattern (IVP) over the sample period. We investigate the validity of this assumption by allowing IVP weights to depend on, for example, the sign of returns from day t-1 or the sign of overnight returns. Estimation results from 5-minute intra-day GARCH for four equity indices during 2019/01/03 - 2021/03/31 show that for those days with negative previous or overnight returns, squared-return-based IVP weights increase significantly in the early morning hours, suggesting a leverage effect or asymmetric IVP. For the jump-robust IVP estimator, a strong asymmetric response is found for days with realized variance increasing over the previous day. Our result is robust against a GARCH-t specification and complements recent studies on time-varying IVP. Price jumps obtained using the proposed IVP adjustment show markedly different distribution over trading hours.