报告题目：Forecasting Volatility with Multi-Horizon Extreme Values: A Heterogeneous Conditional Autoregressive Range Model

报 告 人：周雨田 教授

报告时间：2023年7月6日（周四） 下午15:30

报告地点：经济管理学院C301会议室

报告人简介：

周雨田，台湾中央研究院经济研究所研究员，西安交通大学金禾经济研究中心客座教授，博士生导师。主要从事金融计量，财务经济，宏观经济等领域的研究。1988年博士毕业于加州大学圣迭戈分校(UCSD)，师从诺贝尔经济学奖获得者罗伯特•恩格尔(Robert F. Engle)。先后在Journal of Econometrics, Journal of Money, Credit and Banking, Journal of Banking and Finance, Journal of Applied Econometrics, Journal of Economics Dynamics and Control, Oxford Bulletin of Economics and Statistics等著名国际SSCI期刊发表论文40余篇，其中论文 “ARCH Modelling in Finance”在 Google Scholar有6000多次引用。周雨田教授还先后多次被邀请参加国际重要学术会议，担任多个学术期刊的编辑，近年多次名列经济学名人录“Who’s Who in Economics”。

报告内容简介：

An implied assumption in the conditional autoregressive range (CARR) model is that the data-generating process is short memory, which is not supported by financial data. To resolve the drawback, we construct a new formulation of the CARR process to take into account the range volatilities realized over different time horizons, dubbed the heterogeneous CARR (HCARR) model. In spite of the simplicity of its structure, empirical and simulation results obtained on a broad variety of international equity markets show that the modified HCARR model is able to preserve the persistence inherent in volatility shocks. Importantly, the HCARR model performs significantly better than the short-memory CARR model for both in-sample fit and out-of-sample forecasting ability, especially at longer forecast horizons. We also compare the out-of-sample forecast accuracy of the proposed model with other well-known long memory models. The results reveal that the new model delivers superior forecasts at short horizons up to one month ahead while the dominant performance with longer horizons is market-specific. Finally, our results are quite robust to employ other volatility measures and forecast evaluation methods.

欢迎广大师生参会交流。

经济管理学院

2023年7月5日