Original scientific paper
https://doi.org/10.1080/1331677X.2022.2076144
Intraday volatility analysis of CSI 300 index futures: a dependent functional data method
Danni Wang
Zhifang Su
Qifang Li
Abstract
This study introduces a new volatility model based on dependent
functional data to investigate the intraday volatility characteristics
of CSI 300 in the context of high-frequency data. The volatility
curve is fitted and reconstructed using three methods: functional
principal component analysis, Newey-West kernel, and truncationfree
Bartlett kernel. We adopt a functional time series approach
for short-term dynamic forecasting. The empirical results show
that the proposed dependent functional volatility estimation
model based on the long-term covariance of the truncated
Bartlett kernel can accurately capture the intraday volatility trajectory
and outperforms other models in terms of forecast accuracy
and profitability. This study improves the volatility-related
research methodology, which is conducive to discovering the
price formation mechanism of the stock index futures market and
improving risk management capabilities.
Keywords
Stock index futures; intraday volatility; dependent functional data; functional regression forecasting
Hrčak ID:
303732
URI
Publication date:
31.3.2023.
Visits: 509 *