| 基于非对称Hawkes模型的中国股市价格跳跃建模 |
| Modeling Jumps in Chinese Stock Market Based on Asymmetric Hawkes Mode |
| 摘要点击 37 全文点击 0 投稿时间：2025-06-17 修订日期：2026-02-02 |
| 查看/发表评论 下载PDF阅读器 |
| 中文关键词 非对称Hawkes模型; 高跳现象; 高频交易; 金融市场微观结构; 似然估计 |
| 英文关键词 asymmetric Hawkes model; high jump; high-frequency trading; microstructure of financial market; likelihood estimation |
| 基金项目 国家自然科学基金项目（面上项目，重点项目，重大项目） |
| 投稿方向 金融工程 |
| 作者 | 单位 | 地址 | | 孙铭顺 | 北航经济管理学院 | 北航经济管理学院 | | 杨海军* | 北航经济管理学院 | No.37 Xueyuan Road | | 牟晖 | 北航经济管理学院 | |
|
| 中文摘要 |
| 对称Hawkes模型可以捕捉高频交易中价格的高跳现象，但其对称性假设扭曲了金融资产价格波动的特性. 本文放松了对称Hawkes模型里条件强度的常数项、~维度内影响参数、~ 维度间影响参数、~ 影响衰减参数分别相同的参数假设, 得到了只假设单一维度内影响衰减速度相同的非对称Hawkes理论模型. 随后分别使用对称Hawkes模型和非对称Hawkes模型对上证指数高频交易数据分组识别``高跳" 并建模, 结果表明两个模型均可较好解释高频市场的各种金融现象, 不同数据分组的模型参数相差不大, 稳健性较强; 当市场风险增大时, 即价格变化和波动率增大时, 两个模型的常数项和维度内影响参数都显著增大; 对比两模型参数发现, 非对称Hawkes模型综合参数更能拟合价格变化趋势, 且对称Hawkes模型的参数假设在样本期间均不成立. |
| 英文摘要 |
| The symmetric Hawkes model can capture the ``high jump" in high-frequency trading, but its symmetry assumption distorts the price volatility characteristics of financial assets. In this paper, we relax symmetric Hawkes model’s parameter assumptions about condition intensity, which contain the same constant term, the same parameter of influence within dimension, the same parameter of influence among different dimensions and the same parameter of influences’ attenuation speed, to get theoretical model of asymmetric Hawkes model. The asymmetric Hawkes model only assumes the same attenuation rate within each dimension. Then, we use the symmetric Hawkes model and asymmetric Hawkes model, respectively, to identify the ``high jump" and to conduct empirical research on high-frequency trading data of the Shanghai Securities Composite Index (SSCI) by the group. The results show that both models can well explain various financial phenomena in high-frequency markets, and the model parameters are similar in the case of 240 or 300 data in each group. When market risk increases, i.e., When price volatility and change increase, the constant term and the parameter of influence within dimension of both models increase significantly. By comparing the parameters of the two models, it can be seen that the comprehensive parameter of the asymmetric Hawkes model can fit the price trend better, and all of the symmetric Hawkes model parameter assumptions are not valid during the sample period. |
| 关闭 |
|
|
|
|
|
