趋势持续时间与价格变化相依结构下的高频交易CVaR模型  

High-frequency Trading CVaR Model Based on Dependence Structure of Trend Durations and Price Changes

作　　者：潘水洋[1] 王一鸣[1] 

机构地区：[1]北京大学经济学院

出　　处：《数量经济技术经济研究》2015年第10期139-152,共14页Journal of Quantitative & Technological Economics

摘　　要：针对现有文献估计高频交易风险与实际风险存在偏误,提出基于趋势持续时间与价格变化相依结构下的CVaR模型。该方法首先定义了趋势持续时间和价格变化幅度,并得到趋势持续时间和趋势持续期内价格变化幅度两者边缘分布。然后结合Copula理论构造出趋势持续时间和价格变化幅度的联合分布和条件分布,并在此基础上计算CVaR。最后采用沪深300股指期货高频交易数据对本文提出的模型进行了实证检验。结果表明:下跌趋势持续时间要比上涨趋势持续时间长,对应的下跌幅度要比上涨幅度更大,股指期货上涨与下跌风险具有不对称性。In order to overcome the shortcomings of estimation on the high-frequency trading risk model, a model of CVaR of high-frequency trading based on dependence struc- ture of trend durations and price changes is proposed in this paper. The trend durations and price changes are defined from high frequency data. The type of the marginal probability dis- tribution for the trend durations and price changes is obtained. The copula methods are then employed to determine joint and conditional distribution. The CVaR is calculated from the conditional distribution. An empirical analysis of high frequency data for the stock index fu- tures of China is presented, and the asymmetric property of the risk for rising and falling are verified from the angle of CVaR.

关 键 词：持续时间 连接函数 条件在险价值 

分 类 号：F830.9[经济管理—金融学]