A0888
Title: Stock market volatility forecasting: Can interval data improve it?
Authors: Yongmiao Hong - Cornell University (United States)
Shouyang Wang - Academy of Mathematics and Systems Science, Chinese Academy of Sciences (China)
Meiting Zhu - University of Chinese Academy of Sciences (China) [presenting]
Zishu Cheng - Academy of Mathematics and Systems Science (China)
Jiani Heng - Academy of Mathematics and Systems Science (China)
Abstract: Estimating and forecasting stock volatility is critical to portfolio allocation, asset pricing, and risk management for market participants to make investment decisions and for policymakers to make economic policy. We study the interval-valued models to check whether these models can improve the forecast accuracy in stock volatility, especially the threshold autoregressive interval (TARI) model that utilizes the nonlinear information of the interval-valued stock indexes. Our data sample consists of 19 developed and emerging stock market indices. We find that the TARI model is superior in stock volatility forecasts to other point-based models GARCH and TGARCH models, as well as the conditional autoregressive range (CARR) model proposed based on the range data. As the forecast horizon increases, the difference in forecast accuracy of most market indices becomes statistically indistinguishable. To evaluate the practical implications of our findings, we study a portfolio problem, which reveals that asset allocation based on the interval model forecasts outperforms asset allocation based on other competing models.
