节点文献

基于AH股高频数据下的量化投资策略研究

A Quantitative Investment Strategy Based on High-frequency Data between AH Shares

【作者】 陈振宇

【导师】 欧阳红兵;

【作者基本信息】 华中科技大学 , 金融(专业学位), 2014, 硕士

【摘要】 本文设计了一种基于高频交易下的量化投资策略。选取交叉上市的公司作为研究对象,利用AH股作为样本研究了跨境套利策略的可行性。通过对AH股之间的价差进行建模,利用模型残差构建交易策略进行套利。随着量化投资的兴起,国内目前对于量化策略的研究相对较少。在本文中,考虑到高频交易的特性,选取两地市场流动性比较高的6对股票的价差进行建模,并利用马氏链的转移概率矩阵寻找交易机会。同时,对执行策略时可能产生的成本分为固定成本与可变成本,并对两种成本对收益产生的影响进行详尽的分析。结果显示,利用ARMA模型对价差进行建模后,由于残差转移概率的不同,策略在买入情况下均获得良好净收益,但在卖空情况下则表现为亏损。在对盈利组合进行进一步分析时发现,策略的交易成本大小取决于执行过程中的市场流动性状况。因此,在执行高频交易策略时,必须要考虑到当前市场环境并对交易成本作出精确估计。最后,在综合考虑各种成本的情况下,该策略仍然获得10%的年化收益,并取得了非常良好的表现。因此,对机构投资者来说有非常实际的意义。

【Abstract】 This paper presents a quantitative investment strategy based on high-frequencytrading. Selecting cross-listing companies as a sample, we examine the feasibility ofcross-border arbitrage strategies using AH shares. Specifically, we create a tradingstrategy using residual of the model to arbitrage by modeling the spread between theAH shares.This paper also provides a trading strategy under high-frequency situation andusing Markov chain transition probability matrix to find trading opportunities.Meanwhile, a detailed analysis is made about the transaction costs whenimplementing the trade tactics. The results show that, due to different transitionprobabilities, the strategy gets a net profit in buying circumstance. However, in theshort sale situation, it manifests a loss. In the further analysis of profit group, the scaleof transaction cost relies on market situation in implementation.Finally, considering all transaction costs, this strategy still gets a10%annualized gainand achieves a very good performance. Hence, this strategy has a practical implicationfor institutional investors.

  • 【分类号】F832.51
  • 【被引频次】2
  • 【下载频次】352
节点文献中: